The ATLAS (http://atlas.cern) and CMS collaborations at the Large Hadron Collider (LHC) at CERN discovered the Higgs boson in 2012, a major milestone in our understanding of fundamental particles and their interactions. This discovery led to the 2013 Nobel Prize in Physics awarded to François Englert and Peter Higgs. Since then, the LHC experiments have collected substantially larger datasets during Runs 2 and 3, enabling increasingly precise measurements of the properties of the Higgs boson and extensive searches for new phenomena beyond the Standard Model. Run 3 data-taking was completed in 2026, and the LHC is now in its third Long Shutdown (LS3), during which the accelerator and the experiments are undergoing major upgrades in preparation for the High-Luminosity LHC (HL-LHC).

The HL-LHC is expected to start delivering collisions to the experiments in 2030 and will substantially increase the integrated luminosity available to the ATLAS experiment, with a target of around 3,000 fb?¹ over its physics programme. The ATLAS detector is undergoing a major upgrade to cope with the much higher collision rate and radiation levels expected during this phase. In particular, the current Inner Detector is being replaced by a new all-silicon Inner Tracker (ITk), providing improved tracking performance and extended coverage in the challenging environment of the HL-LHC.

The tracking information provided by the ITk is essential for reconstructing the trajectories of charged particles and for identifying jets originating from the hadronization of bottom quarks. This latter capability, known as b-tagging, is a key ingredient of the ATLAS physics programme. It plays an important role in measurements involving top quarks and the Higgs boson, as well as in searches for new physics and in the study of Higgs boson pair production. In recent years, b-tagging algorithms based on modern machine-learning and deep-learning techniques have led to significant improvements in performance. New generations of these algorithms are now being developed and optimized for the challenging conditions expected at the HL-LHC.

The student will use detailed Monte-Carlo simulations of the upgraded ATLAS detector to study the performance of b-tagging algorithms with the future ITk detector. The project will involve evaluating the impact of the latest detector configuration and reconstruction software on flavour-tagging performance, and investigating possible improvements to the algorithms and their inputs. The project provides an opportunity to gain experience in a broad range of topics, including LHC physics and the Higgs sector; silicon particle detectors; charged-particle tracking and pattern recognition; and modern machine-learning techniques applied to particle physics. The work will make use of existing ATLAS analysis and software frameworks, with a significant part of the analysis based on Python. Prior experience with Python is desirable but not mandatory.

The properties of the Higgs boson have a profound impact on the evolution of the Universe. The operation of the High-Luminosity Large Hadron Collider (HL-LHC), starting in 2030, will provide an unparalleled opportunity to deepen the understanding of the Higgs boson. However, the high pile-up collision environment at the HL-LHC poses significant challenges for data taking and event reconstruction in the ATLAS experiment. To mitigate the effects of high pile-up and fully exploit the physics potential of the HL-LHC, new readout electronics will be installed for the ATLAS liquid argon (LAr) calorimeter. The new electronics will support the reconstruction of calorimeter cell energies using neural networks (NNs). The CPPM ATLAS group plays a leading role not only in the development of the NN algorithms and the upgrade of the LAr calorimeter electronics, but also in studies of Higgs boson physics involving the decay of the Higgs boson into a pair of photons (H->yy).

High-quality reconstruction of LAr cell energies is important for the study of Higgs boson physics in the ATLAS experiment at the HL-LHC. The internship will focus on reconstructing electrons and photons using LAr cell energies reconstructed by NNs. This work may be followed by an evaluation of NN performance using simulated collision events, such as those of the H->yy process. Knowledge and experience in C++/ROOT or Python programming and machine-learning tools would be advantageous, but are not mandatory.

The discovery of the Higgs boson by the ATLAS and CMS collaborations in 2012 completed the particle content of the Standard Model (SM). A fundamental question nevertheless remains open: the shape of the Higgs potential that drives electroweak symmetry breaking. In the SM, this shape is fixed by the Higgs boson self-coupling, which has never been measured. Its value determines the nature of the electroweak phase transition in the early Universe and the stability of the vacuum, and any deviation from the SM prediction would be a clear sign of new physics. At the LHC, the only direct way to measure the self-coupling is through the production of Higgs boson pairs (HH). This process is about a thousand times rarer than single Higgs production and has not yet been observed, which makes it one of the flagship goals of the LHC Run 3 and High-Luminosity programmes. The ATLAS group at CPPM has played a leading role in Higgs boson studies since the discovery and is now a major contributor to the ATLAS HH search.

This internship will focus on the HH->bbyy channel where one Higgs boson decays to two b-quarks and the second decays to two photons. The studies will use simulated samples and the data collected by ATLAS during the run 3 of the LHC. The successful candidate will work within a team of four researchers, on postdoc and two PhD students at CPPM. He/She will analyze the kinematic and topological distributions of the signal in order to improve the selection of signal events and separate them from the background. A machine learning technique, such as neural networks, can also be applied to further improve the signal selection. The significance of the signal will be assessed through a statistical analysis. The student will have the chance to collaborate with other researcher working on the same subject in various international laboratories and present his/her work at CERN.

Prior knowledge of programming languages, especially C++/root or python, is an advantage but is not mandatory.

Since the discovery of the Higgs boson at the LHC in 2012 , understanding its intrinsic properties and interactions has been a priority. The Higgs self-coupling is directly related to the shape of the Higgs scalar field potential, which is important for understanding the mechanism of electroweak symmetry breaking and serves as an essential test of the electroweak theory.

Looking of the production of a pair of Higgs boson is directly sensitive to this self-coupling. The ATLAS team at CPPM has been involved for years in several analyses looking for this di-Higgs process, in particular in the most sensitive decays channels, $bb\gamma\gamma$ and $bb\tau\tau$ . An evidence of this process could be reached with the dataset of the Run 3 of LHC which ended a few months ago.

A complementary approach to study the Higgs boson self-coupling is to use processes in which a single Higgs boson is produced. This is due to the presence of higher order corrections to the single-Higgs cross-section, accounting for a small contribution to the total production of single Higgs of approximately 1-5%. However, considering the single-Higgs production rate is three orders of magnitude higher than the di-Higgs process the contribution of the Higgs self-interaction to single-Higgs measurements is not negligible. Combination of di-Higgs and single-Higgs measurements has been done in the past [1] but the single-Higgs measurements have never been optimised for maximising their complementarity and sensitivity to the relevant Higgs-sector parameters.

This intership will focus on the parameterisation of the single-Higgs cross-section and as a function of the Higgs self-coupling modifier. It will be done using simulated events made privately. If time allows the student will start the optimisation of the selection of events to be sensitive to the Higgs self-couplings in the diphoton channel.

Desired Training: Particle physics, C++/python

Internship Duration (months): 4

Internship Period: March – July

PhD Continuation: planned (if funding is obtained)

Application: please send a detailed CV, a motivation letter and full academic transcripts.

The CNRS promotes the diversity of profiles without distinction of gender, disability or social, ethnic or geographical origins, and wishes to offer the possibility of joining an institution of research excellence to the greatest number.

[1] Constraints on the Higgs boson self-coupling from single- and double-Higgs production with the ATLAS detector using pp collisions at $\sqrt s$ = 13 TeV, ATLAS Collaboration, Phys.Lett.B 843 (2023), 137745, DOI: 10.1016/j.physletb.2023.137745

Decays of heavy-quark hadrons allow to perform indirect searches for effects beyond the Standard Model, by comparing the measured decay properties to their Standard Model predictions. Although our world is made of baryons, baryons containing heavy quarks remain much less well understood than their meson counterparts.

Although the baryons that contain a charm quark and at least one strange quark have been studied for more than 40 years, there are still significant gaps in our knowledge of their properties and decays. This is a significant limitation for any searches for physics beyond the Standard Model involving heavy-quark baryons. Improving our knowledge of these particles is therefore essential in order to turn heavy-baryon decays into precision discovery tools.

The Belle II detector at KEK (Japan) is aimed at precision measurements of properties of beauty and charm hadrons, as well as tau leptons (https://inspirehep.net/literature/1692393). The Belle II collaboration consists of more than 1000 scientists and is taking data since 2019. The key feature of the Belle II detector is the nearly 4pi angular acceptance, which allows to perform the full reconstruction of the visible collision products and calculation of the missing energy. Combined with conservation laws in e+e-->ccbar process (baryon number, electric charge, quark flavours), this allows to fully reconstruct only one charm hadron and deduce the properties of the other one. This allows to study decays previously deemed impossible, such as decays into final states that involve particles difficult to reconstruct (neutrons, neutrinos etc).

During this internship, you will learn experimental techniques used in heavy-flavour physics, notably reconstruction of charged and neutral particles, particle identification, and particle tagging. You will apply your knowledge of relativistic kinematics and conservation laws and learn data analysis techniques. You will work with Belle II data and simulation.

Experience in python/C++ and machine learning techniques is welcome but not mandatory. Curiosity and willingness to learn are more important.

Activities:

Data analysis.

Work context:

The successful candidate will be based at CPPM, Marseille (https://www.cppm.in2p3.fr). The internship will be performed in scope of the ANR StrangeCharm project, where you will collaborate with the researchers involved in this project as well as the other Belle2 group members.

Additional information:

Interested students are invited to submit their application that includes: a motivation letter, curriculum vitae, and grade records. They should also arrange for one letter of recommendation.

Additional references:

https://arxiv.org/abs/1312.7826

https://arxiv.org/abs/1811.09738

https://arxiv.org/abs/1904.12093

The Belle II experiment operates at the SuperKEKB asymmetric-energy $e^+e^-$ collider at KEK, Japan, primarily at the $\Upsilon(4S)$ resonance, which decays almost exclusively into pairs of B mesons. This exceptionally clean and kinematically constrained production environment, in which the initial-state four-momentum is precisely known, makes B factories uniquely well suited to measurements and searches involving missing energy and momentum.

A major experimental challenge is that many of the most interesting B-meson decays contain particles that escape detection. These include neutrinos in semileptonic and rare leptonic decays, as well as potentially invisible particles in searches for physics beyond the Standard Model. To study such decays, Belle II employs B-tagging: reconstructing the other B meson produced in the event, known as the tag side, to determine its four-momentum. Together with the precisely known initial-state four-momentum, this provides a powerful constraint on the signal side, allowing missing particles to be identified through the event's missing energy and momentum.

B-tagging is therefore a cornerstone of the Belle II physics programme, underpinning missing-energy searches as well as precision measurements of semileptonic decays and the determination of the CKM matrix elements $|V_{cb}|$ and $|V_{ub}|$ . The efficiency, purity, and kinematic resolution of the tag directly affect the sensitivity of these measurements.

The CPPM Belle II group is actively developing a new generation of B-tagging techniques to improve tagging efficiency, purity, and kinematic resolution, and to explore their impact on key Belle II measurements and searches. The internship will build on this ongoing effort, connecting the development of advanced B-tagging methods to concrete physics applications involving missing-energy B decays. The student will gain hands-on experience with the full chain of a modern flavor-physics analysis, from the development and validation of reconstruction algorithms to their application in physics measurements. The student will benchmark existing tagging methods, contribute to the development of novel approaches, and study how improvements in tagging performance can enhance sensitivity to semileptonic decays and searches for physics beyond the Standard Model.

Dark matter is one of the main puzzles in fundamental physics today. Indeed, its contribution to the total mass of the Universe would be about 85%, but it cannot be explained in the framework of the Standard Model (SM) of particle physics. Several candidates, however, exist in theories beyond the SM, and the WIMP (weakly interacting massive particle) is one of the best motivated, as it allows to also solve the SM hierarchy problem, directly linked to the stability of the Higgs boson mass.

Experiments searching directly for dark matter thus use the dark matter halo of our galaxy as a potential source of WIMPs. Since 2010, the most sensitive detection technology is based on the measurement of the scintillation light emitted when a WIMP scatters off an atom of noble liquid – argon or xenon. The DarkSide-20k experiment (DS-20k) is currently being installed 1.4 km underground in the Gran Sasso laboratory in Italy. It will use a time projection chamber (TPC) filled with 50 tons of purified argon and read out by 200,000 silicon photomultipliers (SiPMs). This will allow to have the world leading discovery potential for WIMPs after a few years of data taking. The current work is dedicated to the construction of the detector, and data taking should start in 2029. The increase of the liquid argon volume, compared with previous experiments, will allow DS-20k to have the best sensitivity of all liquid argon detectors and competitive sensitivity with respect to xenon detectors, ensuring technological complementarity that is necessary to confirm or infirm potential discovery claims for discovery.

Once light is collected by the SiPMs it must be reconstructed to provide the information on the physical interactions. That step is the link between data and physics analysis. The precision of the energy, position, or event topology reconstruction has direct impact on the signal against background separation, and therefore on the sensitivity to any dark matter candidate. The goal of this internship is to contribute to the improvements of some of the algorithms embedded in the reconstruction chain, and quantify the impact on the projected limits on WIMP-nucleon cross-section. The development of these algorithms can imply innovative techniques such as machine learning approaches and must be implemented in a highly optimised framework to make them suitable for the huge amount of data that will be processed.

This project could be continued as part of a PhD thesis.

More details about the CPPM Dark Matter team:

https://www.cppm.in2p3.fr/web/en/research/astroparticles/#Dark Matter

The Photon team at CPPM participates to the preparation and construction of the CTAO (Cherenkov Telescope Array Observatory). CTAO is a worldwide project to construct the next generation ground based very high energy (VHE) gamma-ray instrument [1]-[2]. VHE gamma-ray astronomy (above 100 GeV) explores the most extreme scenarios in the Universe such as supernova remnants, micro-quasars, gamma ray bursts and active galactic nuclei. CTAO will use tens of Imaging Air Cherenkov Telescopes (IACTs) of three different sizes (Small-Sized Telescope (SST), Medium-Sized Telescope (MST) and Large-Sized Telescope (LST)) distributed across two sites, in Chile and the Canary Island. CTAO will gain a factor of ten in sensitivity with respect to current IACTs and thanks to its performance, it will have the potential to discover and identify hundreds of new sources. The first telescope, LST-1, the prototype for the Northern Site, has been producing scientific data since 2020. The commissioning of three additional LSTs is now beginning, and the full array of telescopes at both sites will be progressively deployed by 2032.

One of the key scientific goals of CTAO is to address the longstanding question of the origin of Galactic very-high-energy cosmic rays by searching for Galactic sources capable of accelerating cosmic rays up to PeV energies, known as PeVatrons.The last few years have been extremely exciting for the PeVatron search since the large-field-of-view LHAASO observatory has detected several ultra-high-energy gamma-ray sources, with photon energies above 100 TeV, proving that PeVatrons exist in our Galaxy [3]-[4]. Nevertheless, due to the poor angular resolution of LHAASO the nature of several detected PeVatron sources remains unknown. While waiting for the completion of CTAO, LST-1 has been observing these intriguing sources for several years, with the main goal of identifying their nature through morphological studies, taking advantage of the much better angular resolution of IACTs compared to LHAASO. Most PeVatron candidates are indeed extended sources, requiring dedicated analysis techniques. The CPPM Photon team has developed recognized expertise in the analysis of extended gamma-ray sources and is leading the analysis of two such candidates.

During the internship, the student will analyze LST-1 data collected on PeVatron candidates during the 2026 observation campaign. The main objective will be to detect and characterize these sources in the 100 GeV–10 TeV energy range and, if possible, to constrain the physical mechanisms responsible for the ultra-high-energy emission observed by LHAASO. The analysis will be performed using the standard analysis tools developed for LST and CTAO, including Gammapy [5]. The analysis framework is based on Python, and the candidate should therefore have a basic knowledge of Python programming.

The internship will provide the student with experience in high-energy astrophysics, gamma-ray data analysis, statistical inference, and scientific programming.

This internship can eventually be continued with a PhD thesis.

[1] Science with the Cherenkov Telescope Array: https://arxiv.org/abs/1709.07997

[2] https://www.cta-observatory.org/

[3] Z. Cao et al. Nature, 594, 33–36 (2021)

[4] Z. Cao et al. (2023) https://arxiv.org/abs/2305.17030

[5] https://docs.gammapy.org/2.1/

The KM3NeT experiment is a next-generation neutrino telescope currently being built in the Mediterranean Sea. It is designed to detect high-energy cosmic neutrinos and to study some of the most energetic phenomena in the Universe. Since neutrinos can travel over cosmological distances without being significantly deflected or absorbed, they provide a unique view of astrophysical environments that are difficult to probe with electromagnetic observations alone. The detection of high-energy neutrinos can therefore provide important information about the physical processes responsible for the acceleration of cosmic particles and contributes to the rapidly developing field of multi-messenger astronomy.

Multi-messenger astronomy relies on the coordinated observation of astrophysical sources using different types of messengers, including neutrinos, gravitational waves, gamma rays, X-rays, and optical or radio emission. The identification of a common source across different messengers can provide much stronger constraints on the nature and physical mechanisms of transient astrophysical events.

The French multi-messenger program involves several complementary instruments and collaborations. Among them, the SVOM mission, a French-Chinese satellite dedicated to the study of transient high-energy phenomena, provides rapid observations in X-rays and gamma rays. COLIBRI is a 1.3-m robotic optical and near-infrared telescope located at the San Pedro Mártir Observatory in Mexico. It is designed for the rapid follow-up of SVOM alerts and other transient sources. The CPPM is involved in the scientific exploitation of these observations and is responsible for the official data-analysis pipeline of COLIBRI.

The combination of KM3NeT neutrino observations with rapid electromagnetic follow-up by instruments such as SVOM and COLIBRI provides an opportunity to search for the astrophysical sources of high-energy neutrinos and to study transient phenomena from a multi-messenger perspective. A particularly important aspect of this effort is the development of real-time alert systems, which can identify promising neutrino candidates and rapidly trigger observations by other observatories.

During this internship, the student will participate in several complementary aspects of multi-messenger research in the KM3NeT context. The work will involve the analysis of neutrino data, the study and development of real-time alert strategies, and the definition and analysis of follow-up observations with electromagnetic observatories.

A first part of the internship will focus on KM3NeT neutrino data and real-time alerts. The student will contribute to the analysis of neutrino candidates and to the identification of events with properties potentially indicating an astrophysical origin. Depending on the progress of the internship, the work may involve studying the criteria used to generate real-time alerts, characterizing the resulting alert sample, and investigating how these alerts can be used to trigger observations by other facilities.

A second part will focus on the multi-messenger follow-up of neutrino candidates. The student will participate in the analysis of electromagnetic observations obtained by SVOM and COLIBRI around the time and position of selected KM3NeT neutrino events. Particular attention may be given to transient sources such as Gamma-Ray Bursts and their afterglows, as well as other variable or transient astrophysical phenomena. The student will investigate methods for identifying spatially and temporally coincident sources and for assessing whether a potential association between neutrino and electromagnetic events is statistically significant.

The internship may also involve the development or improvement of tools used to coordinate and analyse multi-messenger alerts and datasets. This could include Python-based tools for data selection, cross-matching, visualization, statistical analysis, or the characterization of candidate counterparts. Depending on the progress of the work, the student may also contribute to the definition of observing strategies for the rapid follow-up of promising KM3NeT alerts with COLIBRI or other electromagnetic facilities.

The project will make use of real data from the KM3NeT, SVOM and COLIBRI experiments and will provide the student with an opportunity to work on the complete chain from the detection of an astrophysical candidate to its multi-wavelength follow-up and scientific interpretation.

More broadly, this internship will contribute to the development of real-time multi-messenger astronomy, with the goal of identifying the astrophysical sources of high-energy neutrinos and understanding the physical mechanisms responsible for some of the most energetic transient phenomena in the Universe.

This internship can be followed in a PhD.

Required Skills

• Background in astrophysics, particle physics, and/or data science;

• Python programming, ideally at an intermediate to advanced level;

• Experience with other programming languages or software-development tools is a plus;

• Interest in multi-messenger astronomy, neutrino astronomy, and transient astrophysical phenomena;

• Ability to work in an international and collaborative research environment;

• English.

The Integrated Sachs-Wolfe (ISW) effect provides a probe of the late-time evolution of gravitational potentials. Cosmic voids and superclusters represent extreme underdense and overdense environments of the large-scale structure and are expected to leave characteristic imprints on the CMB temperature through the ISW effect.

The aim of this project is to search for the ISW signal associated with cosmic voids and superclusters identified in the Euclid Data Release 1 (DR1), using CMB temperature maps from the Planck satellite. The student will analyse the properties of voids and superclusters identified in Euclid DR1 and investigate their dependence on redshift, size and density.

The main objective will be to measure the average CMB temperature around these structures through a stacking analysis, and to characterise the resulting temperature profiles for voids and superclusters. The analysis will be validated using simulations and null tests, including randomised catalogues and the impact of the Euclid and Planck masks.

Particular attention will be given to the dependence of the measured signal on the properties and selection of the structures, as well as to observational effects such as photometric-redshift uncertainties and survey geometry.

The context: More than twenty years after the discovery of the accelerated nature of the Universe's expansion, there is still no definitive explanation for its physical origin. Several types of dark energy or even alternatives/extensions to general relativity have been proposed in the literature attempting to explain the acceleration of the expansion. By accurately measuring of both the expansion rate of the Universe as well as the growth rate of structures as a function of cosmic time, we can learn more about this cosmological mystery. Particularly at low redshift when the expansion is accelerated and dark energy dominates the expansion, we are interested in obtaining the best constraints on the growth rate of structures. These measurements can be achieved by combining galaxy positions and their velocities. The statistical properties of the density and velocity field are tightly connected to the underlying cosmological model.

Experiments: Measurements of the expansion and growth rates of the Universe are the main scientific goal of current and future experiments such as the Dark Energy Spectroscopic Instrument (DESI), the Zwicky Transient Facility (ZTF), Euclid and the Vera Rubin Observatory Legacy Survey of Space and Time (Rubin-LSST).

DESI is currently measuring the 40 million galaxy positions (with their redshift) and their lower redshift sample will be the most complete to date.

The ZTF survey will discover more than 5 000 type-Ia supernovae, from which we can derive galaxy velocities. Rubin-LSST will increase this number to the hundreds of thousands.

Goal of thesis: The selected candidate will work towards the joint analysis of DESI and ZTF datasets, which contain millions of galaxies and thousands of type-Ia supernovae. The candidate will get familiarised with the physics and the statistics of galaxy clustering, will code their own analysis pipeline, test it on state-of-the-art simulations, and apply it on real data. The measurement of the growth rate of structures using DESI galaxies and peculiar velocities from ZTF supernovae will enable tests of general relativity on cosmic scales. This study is a key project in the roadmap of DESI and ZTF collaborations.

Profile required: The candidate has to have large interest by cosmology, statistics, data analysis and programming (we use mostly python). English proficiency and team work skills are also required.

The ESA Euclid mission, launched in July 2023, aims to understand the origins of the accelerated expansion of the Universe by studying dark energy and dark matter through detailed observations. Onboard, the NISP instrument (Near Infrared Spectrometer and Photometer) produces images that are crucial for cosmological analyses.

The infrared detectors of the NISP instrument are exposed to cosmic radiation, which leaves residual energy deposits in the images. Among these events, the “Snow Balls” stand out due to their unique shape: they affect a relatively large area of pixels, have an unusual energy distribution, and appear as small round balls, hence their name.

Unlike other instrumental artifacts, Snow Balls do not directly affect scientific analysis. However, their occurrence rate and evolution over time are important parameters to monitor to ensure accurate tracking of detector behavior.

Internship Objectives:

The internship aims to identify, catalog, and characterize “Snow Ball” events in order to assess their frequency and evolution. The main tasks will include:

- Visual identification and catalog creation: manually annotating Snow Ball events in images to create a reference database.

- Machine learning for detection: training and validating an algorithm for automatic detection and classification using the catalog.

- Statistical analysis: studying the properties of Snow Balls, including:

temporal evolution,

occurrence rate per detector,

correlation between size and deposited energy,

comparison with other cosmic-ray-induced events.

Skills Developed:

- Annotation and processing of astronomical images.

- Development of detection and classification algorithms (Python, ML).

- Statistical analysis of instrumental phenomena.

- Understanding of infrared detector behavior under cosmic radiation.

Twenty years after the discovery of the accelerating expansion of the universe through supernova measurements, the supernova probe remains one of the most accurate means of measuring the cosmological parameters of this recent period in the history of our universe, dominated by the so-called dark energy.

The Rubin Observatory with the Large Survey of Space and Time (Rubin/LSST) will be commissioned in 2025 and will be fully operational by the end of 2025. It is an 8.4-m telescope equipped with a 3.2-billion-pixel camera, the most powerful ever built.

This telescope will take a picture of half the sky every three nights for ten years. This survey will make it possible to measure billions of galaxies with great precision, and to track the variation over time of all transient objects. Together with many other astrophysical studies, it will be a very powerful machine for determining cosmological parameters using many different probes and, in particular, will impose strong constraints on the nature of dark energy. The LSST project aims to discover up to half a million supernovae. This improvement of two to three statistical orders of magnitude over the current data set will enable precise testing of the parameters of dark energy, test general relativity and also impose new constraints on the isotropy of the universe.

In this Master 2 internship, we propose to analyse the first Rubin/LSST images using deep learning methods to detect and identify transients and supernovae.

The LSST group at CPPM is already involved in precision photometry for LSST, with direct involvement in algorithm validation within DESC/LSST [1][2], and has proposed new deep learning methods to improve detection [5], photometric supernova identification [3] and photometric redshifts [4].

[1] https://www.lsst.org/content/lsst-science-drivers-reference-design-and-anticipated-data-products

[2] https://arxiv.org/abs/1211.0310

[3] https://arxiv.org/abs/1901.01298

[4] https://arxiv.org/abs/1806.06607

[5] https://arxiv.org/abs/2607.05393

Transient astronomy has developed considerably in recent years with the emergence of new instruments capable of rapidly detecting and following variable astrophysical phenomena. Among these, Gamma-Ray Bursts (GRBs) are among the most energetic explosions in the Universe. These extremely luminous and transient events are associated with the formation of a compact object and the emission of relativistic jets. They can be produced, in particular, by the gravitational collapse of massive stars or by the merger of two neutron stars.

GRBs exhibit a prompt emission, primarily observed at high energies, followed by a multi-wavelength afterglow that can be detected for several hours or even days. Rapid and regular observations of this afterglow make it possible to characterize the physics of the explosion, the geometry of relativistic jets, and the environment of the sources. Neutron-star mergers are also prime targets for multi-messenger astronomy, as they can be accompanied by both gravitational-wave emission and electromagnetic counterparts, such as short GRBs and kilonovae.

In this context, COLIBRI is a 1.3-m robotic telescope dedicated, among other purposes, to the rapid follow-up of transient sources and high-energy alerts. Located at the San Pedro Mártir Observatory in Mexico, COLIBRI can rapidly obtain photometric observations in several optical and near-infrared bands. Its ability to observe a source shortly after an alert makes it particularly well suited to the study of GRB afterglows and the search for electromagnetic counterparts of transient events.

During this internship, the student will participate in the study and optimization of observing strategies with COLIBRI for the follow-up of transient sources. The main objective will be to investigate how observations can be optimized to maximize the chances of detecting and characterizing electromagnetic counterparts associated with astrophysical alerts. Part of the work will also focus on exploiting Legacy Survey of Space and Time (LSST) alert data to identify and select transients with properties compatible with phenomena such as GRB afterglows. The student will therefore investigate the criteria that can be used to select the most relevant alerts for rapid follow-up with COLIBRI and assess the impact of this selection on the efficiency of the observing strategy.

The work may rely on simulations of source populations and on the instrumental characteristics of COLIBRI. Particular attention may be given to the expected light curves of GRB afterglows and kilonovae, as well as to their spectral and temporal evolution. The student may investigate the influence of various parameters of the observing strategy, such as the delay between the alert and the start of observations, the duration and number of exposures, filter selection, and follow-up cadence.

Depending on the progress of the internship, the work may also make use of real observations obtained with COLIBRI and of analysis tools developed for their photometric processing. This will provide an opportunity to connect simulations, the design of observing strategies, and the scientific exploitation of observational data.

More broadly, this work contributes to the development of transient and multi-messenger astronomy, with the aim of improving our understanding of some of the most energetic astrophysical phenomena in the Universe and optimizing the observational response to new alerts.

This internship can be followed by a thesis.

## Required skills

* Python programming, ideally at an intermediate to advanced level;

* An interest in scientific data analysis and processing;

* English or French proficiency.

Inclusive decays such as $B \rightarrow D X$, $B \rightarrow \Lambda_{c} X$ et $B \rightarrow \D_{s} X$ provide a powerful probe of charm production and the underlying dynamics of B-meson decays. Existing measurements focus on the reconstructed charm hadron, while the accompanying system X is treated inclusively. However, the properties of X contain additional information about the decay process and the hadronization of the quarks produced in the weak decay.

This project will investigate the composition and properties of the X system using Belle~II data and the existing inclusive charm-hadron samples. The study will explore the presence of additional particles and hadrons in X, their kinematic properties, multiplicities, and correlations with the reconstructed charm hadron. These studies can provide complementary information on the underlying decay mechanisms and help identify specific contributions to inclusive B-meson decays.

The CPPM Belle~II group is actively involved in the development of B-tagging techniques and in measurements of inclusive B-meson decays. The internship will allow the student to build on this ongoing effort and gain hands-on experience with Belle~II data, B-tagging, event reconstruction, and inclusive flavor-physics analyses. The project will establish the foundations for a broader program of studying the structure of inclusive B decays and the dynamics of charm production at Belle II.

Context

The study of neutrino properties is currently one of the most active fields in particle physics. Many fundamental questions remain open, such as the origin of their mass or their hierarchy. Experimentally studying neutrinos remains particularly challenging: neutrinos interact extremely weakly with matter, and their characteristics are currently inferred only from the rare interactions observed in large detectors. Measurements of individual neutrino properties in these detectors are still limited in precision — their energy, for example, is generally known only to within 10 to 20%.

To overcome these limitations, researchers at CPPM, in collaboration with several international teams, are developing a new experimental approach called neutrino tagging. This method aims to determine the properties of a neutrino at the moment of its production, during the decay of a pion into a muon and a neutrino. Since the pion and muon are charged particles, their trajectories can be measured with great precision. The characteristics of the neutrino can then be inferred using simple kinematic relations, allowing in particular for an estimate of its energy with a relative precision better than one percent.

The main challenge of this method lies in the extremely high intensity of the pion beams used to produce neutrinos (on the order of 10¹⁰ pions per second). Reconstructing the trajectories of these particles and their decay products represents a major technological and algorithmic challenge.

Internship objective

The internship aims to contribute to the development and evaluation of machine-learning algorithms designed to meet this challenge, as well as to the optimization of the specifications of the trackers required to implement the method.

Skills

• Neutrino physics
• Detector physics (trackers)
• Machine learning
• C++ and Python programming
• LaTeX

Application

We welcome applications from both M1 and undergrade (L3) students.

Please fill out this

https://www.cppm.in2p3.fr/~mperrint/Recruit/apply.php>form using job offer reference M1-26-27-KM-01.

Dark matter is one of the great enigmas of fundamental physics today. Its contribution to the total mass of the Universe is 85%, but it cannot be explained within the framework of the Standard Model of particle physics (SM). However, several candidates for dark matter exist in theories beyond the SM: this is the case of the axion, one of the best-motivated candidates as it also explains the absence of CP violation in the strong interaction.

Detecting axions requires new experimental techniques, and few proposed approaches can cover the theoretically motivated 100 micro-eV mass range. MADMAX (MAgnetised Disk Mirror Axion eXperiment) ultimately aims to address this challenge using the novel concept of a dielectric haloscope, consisting of a series of 80, 1-meter diameter dielectric disks placed inside a 10 Tesla hollow-bore magnet under cryogenic conditions (~10 Kelvin). By precisely controlling the positions of the disks at micrometre precision, the electromagnetic signal produced by axions can be enhanced, making it possible to search for these extremely weak signals of order 10^(-27) Wm^(-2). In 2025, the collaboration reported the first axion search using a dielectric haloscope at CERN, setting world-leading limits around 80 micro-eV and is currently developing and testing new prototype systems, with cryogenic tests and data acquisition planned for 2027.

The student will analyse real microwave measurement data from the current prototype to characterise its electromagnetic response and extract a boost factor: a key quantity describing the detector’s enhancement of an axion-induced electromagnetic signal. The student will gain hands-on experience with microwave measurement techniques, data analysis, and electromagnetic simulations, working as part of an active international collaboration currently in the final phase of prototype testing.

This project is a solid introduction to experimental astroparticle physics, depending on progress and data availability, there may be scope to extend the analysis to future prototypes or to other MADMAX searches, such as dark photons.

[Suggested background: familiarity with Python and classical electromagnetism with an interest in particle physics and cosmology. No prior RF/microwave experience required.]

More details:

https://www.cppm.in2p3.fr/web/en/research/astroparticles/#Dark%20Matter

https://journals.aps.org/prl/abstract/10.1103/c749-419q

https://arxiv.org/abs/2109.07376

Transient astronomy has considerably developed in recent years with the emergence of new instruments capable of rapidly detecting and following variable astrophysical phenomena. Among these, Gamma-Ray Bursts (GRBs) are among the most energetic explosions in the Universe. These extremely luminous and transient events are associated with the formation of a compact object and the production of relativistic jets. They can notably be produced by the gravitational collapse of massive stars or by the merger of two neutron stars.

GRBs exhibit a prompt emission, primarily observed at high energies, followed by a multi-wavelength afterglow that can be detected for several hours or even days. Rapid and regular observations of this afterglow provide important information about the physics of the explosion, the geometry of relativistic jets, and the environment of the source. Neutron-star mergers are also key targets for multi-messenger astronomy, as they can produce both gravitational-wave emission and electromagnetic counterparts, such as short GRBs and kilonovae.

In the context of time-domain astronomy, the Legacy Survey of Space and Time (LSST) will produce a very large number of alerts associated with variable and transient astrophysical sources. The rapid identification and characterization of these sources are essential for selecting the most interesting events for follow-up observations with dedicated telescopes.

COLIBRI is a 1.3-m robotic telescope located at the San Pedro Mártir Observatory in Mexico. It is dedicated, among other scientific objectives, to the rapid follow-up of transient sources and provides photometric observations in several optical and near-infrared bands. Its ability to respond rapidly to alerts makes it particularly well suited to the study of transient phenomena such as gamma-ray burst (GRB) afterglows and electromagnetic counterparts of multi-messenger events.

The goal of this internship will be to participate in the analysis of one or a few transient sources detected by LSST and followed up with COLIBRI. The student will use existing photometric analysis tools developed within the team to exploit the available observations and construct light curves and spectral energy distributions of the sources under study.

The first part of the work will focus on the analysis of LSST and COLIBRI data. This will include extracting and validating photometric measurements, identifying possible detections in different bands, constructing light curves, and studying the evolution of the sources at different times. The student will also use existing tools for forced photometry in order to characterize faint sources.

The second part of the internship will focus on developing simple tools for transient classification. The objective will be to investigate whether the temporal and spectral properties of a source are compatible with different astrophysical scenarios, in particular a GRB afterglow or a kilonova. The student may explore indicators such as the evolution of the source brightness with time, its colours, the slopes of its light curves, and their spectral evolution. These properties can then be compared with models or with samples of previously identified transients.

Depending on the selected sources and the available data, the analysis may also be complemented with information from other astronomical surveys or additional follow-up observations. The goal will not be to develop a complete classifier for the full LSST alert stream, but rather to develop and test a first classification approach on a small number of representative case studies.

This internship will provide the student with hands-on experience in the analysis of real astronomical data, transient-source photometry, multi-instrument observations, and the development of analysis tools in Python. It will also provide an introduction to the scientific exploitation of LSST alerts in the context of time-domain and multi-messenger astronomy.

Required Skills

• Python programming, ideally at an intermediate to advanced level;

• Interest in scientific data analysis and processing;

• English or French.

DEPHY est un projet de recherche et développement soutenu par l'institut IN2P3, qui vise à étudier les technologies nécessaires au développement de détecteurs à petits pixels pour les détecteurs de traces dans les futurs accélérateurs de particules et en particulier le FCC (Futur Circular Collider) au CERN. Parmi les objectifs, le projet se concentre sur la conception de circuits de lecture de pixels à haute résolution temporelle de l’ordre de quelques picosecondes, capables de fonctionner dans des environnements soumis à des rayonnements extrêmes.

Les difficultés sont diverses~: Les applications en physique des particules génèrent un flux de données très important et exigent un fonctionnement des détecteurs dans un environnement sévèrement radioactif. La dernière production d’ASIC «~Front-End~» est tolérante à une dose totale (TID) de 500~Mrad (5~MGray) représentant 5 ans d’exploitation.

Pour les applications où le temps est un facteur critique, la technologie CMOS 28~nm présente un intérêt particulier. Un premier ASIC prototype a été conçu au CPPM avec cette technologie et est en cours de tests et de caractérisation. Cet ASIC comprend différents sous-circuits permettant d’une part de caractériser une matrice de 400 pixels (taille~: 25×25~µm) et d’autre part de déterminer la tolérance de la technologie à la dose ionisante (TID) ainsi qu’aux effets des évènements singuliers (SEE).

Activité principale :

Les tests des ASIC’s aboutissent à la validation des architectures proposées auprès des collaborations en place où les résultats sont présentés. Un banc de tests polyvalent est déployé actuellement au CPPM permettant à la fois de caractériser la partie analogique des pixels (Amplificateur de charge (BW, bruit…)) et de tester en irradiation, sur sites dédiés, certaines parties du circuit. La carte principale se veut compacte et portable. Elle comprend une électronique embarquée de type BeagleBone (OS Linux), communicant avec un FPGA (Altera-Cyclone III - programmation VHDL). Ce dernier gère les signaux nécessaires au fonctionnement de l’ASIC en test (DUT). Le contrôle-commande s’effectue au niveau de la carte BeagleBone en C++. D’autres paramètres tels que, la consommation, la température, les niveaux d’alimentation, sont enregistrés via un bus I2C. Ces éléments sont indispensables pour s’assurer du bon fonctionnement du DUT.

Le stage de 6 mois devra comporter plusieurs étapes~:

• Prise en main du banc de test.

• Maîtrise, débogage des différentes fonctions du banc.

• Amélioration et finalisation de l’ensemble

• Possibilité de proposer une interface utilisateur de type Qt Python.

Connaissances requises :

Le(la) candidat(e) devra posséder :

• des bonnes bases en électronique

• des solides connaissances en programmation VHDL, C++, Qt Python.

Contact :

CV + lettre de motivation avec la référence « Ingenieur-2627-EL-04 » à

Frédéric HACHON

Ingénieur de Recherche - Correspondant des stages techniques du CPPM

Tél : +33 4 91 82 76 71- Mél : hachon@cppm.in2p3.fr

Le stage de 6 mois sera conventionné et rémunéré.

Le stage se déroule au Centre de Physique des Particules de Marseille (CPPM). C’est est une unité mixte de recherche (UMR 7346) qui relève de l’IN2P3, institut regroupant les activités de physique des particules et de physique nucléaire au sein du CNRS et d’Aix-Marseille Université.

Le CPPM étudie actuellement la pertinence de l’utilisation d’une technologie BiCMOS 130 nm dans le but de concevoir un détecteur pixelisé permettant la mesure et la caractérisation d’une particule élémentaire appelée neutrino dont sa masse est extrêmement faible et ne possédant aucune charge électrique.

Les deux principaux chalenges techniques de ce projet est de maintenir les informations temporelles extrêmement précises (< 40ps) tout en étant tolérant à une dose ionisante très élevée de l’ordre de la dizaine de Méga Gray. La pertinence de ce future détecteur réside dans la détection spatiale d’une particule ionisante dont la désintégration amène la création d’un neutrino, particule pas totalement comprise par la communauté des physiciens du domaine. Actuellement, ce futur concept n’a aucune concurrence, et pourrait entraîner des réponses sur les propriétés encore manquantes et donc mieux comprendre son origine et son comportement.

Activité principale :

Le/la stagiaire va rejoindre l’équipe de développement du CPPM qui doit réaliser un premier prototype pour l’année 2027. Le circuit sera réalisé en technologie BiCMOS 130nm et contiendra des fonctions analogiques et digitales ( préamplificateur, des DACs de control, BandGap ou encore des fonctions numériques de contrôle, ou de lecture des données)

Dans un premier temps, le/la stagiaire doit mener une recherche bibliographique détaillée sur les détecteurs à pixels monolithiques et sur les fonctions générales. Ensuite, il lui sera proposé de participer à la conception du circuit prototype selon le cahier des charges fourni.

• Etude bibliographique sur les architectures de la fonction.

• Conception, simulation sous Cadence

• Dessin des masques (Layout)

• Simulation post-layout

Connaissances requises :

• Bonnes connaissances en conception de circuits intégrés en technologie CMOS

Contact :

CV + lettre de motivation avec la référence « Ingenieur-2627-EL-03 » à

Frédéric HACHON

Ingénieur de Recherche - Correspondant des stages techniques du CPPM

Tél : +33 4 91 82 76 71- Mél : hachon@cppm.in2p3.fr

Le stage de 6 mois sera conventionné et rémunéré.

Le Centre de Physique des Particules de Marseille (CPPM), est l’un des laboratoires fédérant les moyens de la Physique des Particules en France. Ses équipes de chercheurs et ingénieurs s’investissent au sein des collaborations internationales des expériences de Physique des Particules, d’Astroparticules et de cosmologie. Elles travaillent, notamment, sur les expériences installées autour du plus puissant accélérateur de particules au monde, le LHC situé au CERN à Genève (http://www.cern.ch).

Vous travaillerez au sein du service électronique, qui a notamment contribué aux systèmes d’acquisition des expériences ATLAS et LHCb du CERN. Ces systèmes traitent des flux de données alant jusqu’à 30~TB/s sur plus de 10 000 liens sériels optiques. Notre équipe a développé une carte d’acquisition prototype basée sur un FPGA Agilex 7 M-series doté de transceivers F-tiles à 112 Gb/s permettant d’atteindre 400 Gb/s de débit réseau. Un premier firmware FPGA intégrant les couches Ethernet, IP et UDP est opérationnel.

La prochaine génération de systèmes vise les transferts Remote Direct Memory Access (RDMA), qui permettent l’échange efficace entre les nœuds dédiés à la lecture des détecteurs et ceux dédiés au traitement en temps réel des données, tels que des GPU. Le protocole RoCE~v2 (RDMA over Converged Ethernet) permet ces échanges à travers un réseau IP. Vous mettrez en œuvre ce protocole sur des SoC FPGA de dernière génération en vue des futures expériences de physique fondamentale.

Activité principale :

Le/la stagiaire rejoindra l’équipe de développement, composée d’électroniciens et d’informaticiens. Sa mission sera de mettre en œuvre une preuve de concept d’un lien 400~Gb/s sur protocole RoCE~v2 entre un FPGA Agilex 7 et une carte d’interface réseau (NIC).

En premier lieu, le/la stagiaire devra mener une recherche bibliographique sur les concepts de réseaux et protocoles utilisés. Il s’appuiera ensuite sur des développements existants de firmware FPGA capables de supporter le protocole UDP. Finalement, le développement de logiciel bas niveau sera nécessaire pour la récolte de données en mémoire du serveur, voire de GPU.

Profil recherché :

Vous préparez un diplôme d’ingénieur dans le domaine de l’électronique, des systèmes embarqués et/ou informatique. Les compétences suivantes seront appréciées, et une formation sur les outils utilisés sera fournie~:

• Maîtrise des langages VHDL/Verilog pour la conception FPGA

• Connaissance de l'architecture des protocoles réseau Ethernet

• Familiarité avec les outils de synthèse et simulation FPGA (Quartus Prime, ModelSim)

• Une expérience avec les transceivers série à haut débit (SerDes) est un plus.

• Conception logicielle, langage Python (Polars, Pytest), C++

• Utilisation de la bibliothèque logiciel DPDK

Contact :

CV + lettre de motivation avec la référence « Ingenieur-2627-EL-02 » à

Frédéric HACHON

Ingénieur de Recherche - Correspondant des stages techniques du CPPM

Tél : +33 4 91 82 76 71- Mél : hachon@cppm.in2p3.fr

Le stage de 6 mois sera conventionné et rémunéré.

Le Centre de Physique des Particules de Marseille (CPPM), unité mixte CNRS / Aix-Marseille Université, est l'un des laboratoires de l'Institut National de Physique Nucléaire et de Physique des Particules (IN2P3), institut du CNRS fédérant les moyens de la Physique des Particules en France. Il intervient notamment sur les expériences LHCb et ATLAS installées auprès du LHC, l'accélérateur et collisionneur proton-proton le plus puissant au monde, situé au CERN à Genève (http://www.cern.ch).

Le CPPM conçoit et livre au CERN des systèmes d'acquisition en temps réel à base de FPGA pour ces deux expériences. Ces systèmes assurent la lecture de l'ensemble des données (à 30 Terabits/s) des sous-détecteurs via plusieurs milliers de fibres optiques à 10 Gbits/s, leur traitement en temps réel, puis leur envoi vers une ferme de calcul composée de plusieurs milliers de processeurs.

Fort de son expertise, le CPPM souhaite aujourd'hui évaluer de nouvelles technologies d'alimentations pour répondre à des contraintes en nombre de rails et de consommation toujours plus importantes des FPGA de nouvelle génération. Cette évaluation doit notamment garantir la conservation d’une électronique compacte, fiable et générique à plusieurs familles de FPGA.

Activité principale :

Le/la stagiaire rejoindra l’équipe de développement du CPPM. Sa mission sera de concevoir et de valider un démonstrateur complet de gestion d’alimentation pour FPGA haut de gamme, tel qu’un Agilex 7 du constructeur ALTERA.

Le stage s'organisera autour de deux activités distinctes~:

• l’évaluation de nouvelles technologies d’alimentation pour valider l’aptitude des régulateurs à fournir plus d’une centaine d’ampères à moins de 1V avec une ondulation réduite de l’ordre de quelques millivolts.

• La conception et la réalisation d’un séquenceur d’alimentation prototype sur FPGA ou sur micro-contrôleur. Le contrôleur retenu devra également être capable de configurer ces régulateurs et d’assurer leur surveillance.

Le/la stagiaire sera ainsi amené à travailler à la frontière entre électronique de puissance, étude de FPGA et de micro-contrôleur, firmware et logiciel temps réel, depuis la caractérisation de composants jusqu’à la validation expérimentale sur banc de test.

Profil recherché :

Vous préparez un diplôme d’ingénieur dans le domaine de l’électronique et de l’informatique. Les compétences suivantes seront appréciées, et une formation sur les outils utilisés sera fournie~:

• Utilisation des appareils de mesure~:

oscilloscope, analyseur de spectres~;

• Électronique analogique et numérique~;

• Conception de firmware FPGA en langage VHDL~;

• Conception logicielle RTOS, langage Python et C

Contact :

CV + lettre de motivation avec la référence « Ingenieur-2627-EL-01 » à

Frédéric HACHON

Ingénieur de Recherche - Correspondant des stages techniques du CPPM

Tél : +33 4 91 82 76 71- Mél : hachon@cppm.in2p3.fr

Le stage de 6 mois sera conventionné et rémunéré.

Le CPPM travaille notamment sur l’expérience ATLAS basée au CERN à Genève. Cette collaboration internationale comprend 3 000 scientifiques issus de 174 instituts, représentant pas moins de 38 pays. Une de ses missions est notamment d’œuvrer pour l’horizon 2029 à une mise à niveau de l’électronique de lecture du calorimètre à argon liquide du détecteur ATLAS.

Une équipe d’Ingénieurs et physiciens du CPPM travaille sur la réalisation d’un tout nouveau système d’acquisition de données et de trigger appelé le processeur LASP. Ce projet consiste à développer une carte prototype au format ATCA à base de deux FPGA INTEL Agilex et un contrôleur MAX10 ainsi que d’une vingtaine de modules optiques. Cette carte calculera en temps réel les énergies déposées dans le calorimètre suite aux collisions dans le LHC. L’un des défis de cette carte est le fait qu’elle doive traiter une énorme quantité de données (500 Gb/s par FPGA), tout en calculant l’énergie en moins de 125 ns.

Activité principale :

Il s’agit de développer un firmware comportant l’implémentation d’un réseau de neurones en VHDL sur des FPGAs AGILEX. Ce firmware utilisera l’intelligence artificielle pour calculer les énergies déposées dans le calorimètre à partir de signaux transmis par ses capteurs. Les réseaux de neurones sont déjà développés en software par des informaticiens et des physiciens et des premières versions ont été portées sur le FPGA. Il s’agit maintenant de porter une nouvelle version en firmware VHDL de la façon la plus optimale possible. Ces réseaux nécessitant un grand nombre d’opérations mathématiques, il faut pouvoir les intégrer dans les FPGAs aux ressources limitées et réduire leur temps de calcul à moins de 125 ns. Il s’agit certes d’un défi pour les FPGAs, mais qui est réalisable. De plus, un framework Python de génération de code VHDL à partir du réseau TensorFlow devra être mis à jour en intégrant de nouvelles fonctionnalités.

Les développements et les activités proposés pourront concerner :

• La réalisation d’un firmware comportant un réseau de neurones

• La simulation de ce firmware sous Modelsim pour valider les fonctionnalités

• L’optimisation de l’implémentation sur FPGA pour réduire les ressources et la latence

• Le développement d’un framework python de génération de code VHDL

• La validation de ce firmware de test sur un kit de développement AGILEX et sur carte

Le ou la stagiaire sera accueilli(e) au sein du service informatique qui possède un savoir-faire étendu dans le développement de firmwares FPGA de dernière génération. L’étudiant(e) devra collaborer avec le service électronique qui réalise la carte L’étudiant(e) bénéficiera du matériel requis pour la réalisation complète du firmware et des tests. Des déplacements au CERN à Genève seront possibles pour présenter ses travaux et participer à des réunions de la collaboration.

Connaissances requises et appréciées :

• Développement de firmwares FPGA en VHDL avec Quartus et simulations Modelsim

• Bonnes bases en électronique numérique et en langage Python

• Anglais pour les présentations, la documentation technique et la rédaction de rapports

Contact :

CV + lettre de motivation avec la référence « Ingenieur-2627-IN-02 » à

Frédéric HACHON

Ingénieur de Recherche - Correspondant stages techniques du CPPM

Tél : +33 4 91 82 76 71- Mél : hachon@cppm.in2p3.fr

Le stage de 6 mois sera conventionné et rémunéré.

Context

The laboratory has an IT and network infrastructure, as well as various monitored technical systems and equipment, including network switches, servers, IP devices, air-conditioning units, chillers, temperature sensors, and other related equipment.

The existing monitoring tools already collect various types of information, such as equipment status, response times, temperatures, network errors, and interface usage.

However, monitoring based solely on fixed thresholds can generate a large number of irrelevant alerts or, conversely, fail to detect unusual changes in system behavior.

The objective of the internship will be to implement an initial layer of intelligent anomaly detection capable of analyzing several measurements before generating an alert.

The aim is not to develop a complex artificial intelligence system, but rather to build a simple and explainable system based on the following principle:

Measurement ? history ? detection ? correlation ? alert ? diagnostic assistance

Main Objective of the Internship

Design and develop a prototype capable of automatically detecting certain anomalies within the IT infrastructure, network infrastructure, and/or technical equipment.

The system should notably attempt to distinguish between:

an isolated event with no significant impact;

a persistent anomaly;

several symptoms that may share a common cause.

Particular attention will be paid to reducing false alerts.

Proposed Tasks

1. Understanding the Infrastructure and Monitoring Environment

The intern will first become familiar with:

the network architecture;

the main IT equipment;

the monitored technical equipment;

the monitoring tools used within the organization;

the different types of available data.

The intern may also identify information accessible through:

SNMP;

ping;

system logs;

APIs;

monitoring databases;

temperature sensors.

2. Data Collection and Analysis

The intern will collect selected data from the monitoring system in order to analyze it.

3. Implementation of Detection Rules

The intern will develop several rules designed to detect abnormal situations.

Prototype Development

The intern will develop a small tool capable of analyzing the collected data.

Ideally, the program should be able to:

collect measurements;

maintain a historical record;

apply detection rules;

correlate multiple events;

generate clear and understandable alerts.

System Testing

In order to validate the prototype, different incidents may be simulated or intentionally triggered within a test environment.

Euclid (http://www.euclid-ec.org) is a major ESA mission. This space telescope, dedicated to understanding the Universe, was launched in July 2023 and will map the entire sky. It will provide a 3D-mapping of galaxies with unprecedented precision. These measurements of the distant Universe large structures will test the cosmological model, particularly questioning the nature of dark energy. The mapping will be achieved using the NISP spectrophotometer and its 16 infrared detectors, which were calibrated on ground by CPPM, a fundamental step to validate the instrument's performance.

SVOM (https://www.svom.eu/) is a franco-chinese mission dedicated to the observation gamma-ray bursts, phenomena among the most energetic in the Universe. The Cagire camera will provide a unique ground follow-up in the infrared band thanks to a prototype CMOS IR detector that will be tested for the first time on the sky in 2027.

$\textbf{Main Activities}$

At the cutting edge of technology, the IR CMOS detectors studied at CPPM have already revealed, during their calibration on the PICA platform, a level of persistence which proves to already be a real challenge for the extraction of sky data. A better understanding of this complex effect is necessary. As a first step, we focused on analyzing the ground calibration data. This allowed us to highlight certain dependencies—on temperature, source signal amplitude, and wavelength—that we wish to explore in greater detail. We are therefore preparing for a new round of tests dedicated to these in-depth studies of persistence.

The intern will also be involved in these new tests. They will be responsible for operating the test bench and participating in data acquisition. They will focus on test data and seek to characterize persistence and its counterpart anti-persistence. For these analyses, they will be able to rely on previously developed Python scripts and will aim to:

1) Adapt existing codes to meet new analysis requirements

2) Extract relevant data related to persistence

3) Derive models for the dependencies of persistence and anti-persistence

The intern will participate in team meetings and will be expected to present their work.

$\textbf{Required Knowledge and Abilities}$

- Strong foundation in programming in Python

- Solid understanding of semiconductor physics and signal processing

- Knowledge of statistics is useful

- Interest in experimentation

- Enjoys team work

The 6-month internship will be conventionally recognized and paid. It may be continued as a PhD thesis with CNES funding (application in progress).

$\textbf{Contact:}$ CV + cover letter

Aurélia Secroun, Research Engineer?Tel : 04 91 82 72 15 mail : secroun@cppm.in2p3.fr

Find this offer at https://www.cppm.in2p3.fr/web/fr/emplois_stages/stages/