What you'll do
Join us to push the boundaries of real-time Machine Learning (ML) in one of the most demanding computing environments in the world. You will develop cutting-edge ML models for the CMS Level-1 Trigger – an ultra-low-latency, FPGA-based system responsible for selecting the most interesting LHC collisions in real-time. You will help design the next generation of trigger algorithms for the High Luminosity LHC era by co-training ML models across different systems to maximise physics performance while optimising information flow, bandwidth, and on-device resource usage. This includes developing and scaling MLOps workflows, integrating ML models into FPGAs, and delivering demonstrators that validate full-chain performance from training and physics performance to on-hardware deployment. This position is part of the NextGen Triggers (NGT) project, a 5-year collaboration between LHC experiments and the CERN Research and Computing Departments.
The project leverages innovative Artificial Intelligence technologies and high-performance computing architectures to enhance trigger selection, data processing, and theoretical interpretation for LHC experiments. The insights gained will inform future detector development, data flows, and theoretical tools.
Your responsibilities
- Design and train ML models to boost the physics selections of the CMS Phase-2 Level-1 Trigger by targeting specific signatures and optimising information transport across the multi-algorithm system.
- Develop, deliver, integrate, and test ML models in FPGAs (including RTL/HLS components and software emulators).
- Demonstrate physics performance gains and present results within CMS, at CERN, and at international conferences.
- Design and incorporate MLOps practises, scaling up workflows to ensure reproducible training, validation and deployment of ML-based trigger algorithms.
- Collaborate closely with colleagues in CMS, CERN departments, and external institutes working on ML-for-Trigger research.
Still here? Let's make a quick check about
Your profile
- Experience developing and applying Machine Learning algorithms for physics or scientific data analysis.
- Familiarity with Fast ML / hardware-constrained ML techniques is an advantage.
- Knowledge of physics analysis or physics event reconstruction methods.
- Experience with Trigger and Data Acquisition systems, including hardware architectures.
- Practical experience with software development (e.g. GitHub/GitLab, Continuous Integration, MLOps).
- Basic knowledge of FPGA design including HDLs (VHDL/Verilog) and/or High Level Synthesis (C++).
- Your studies focused on Physics or a related field.
Your skills
- Machine Learning & Fast Machine Learning;
- Physics Data Analysis & Reconstruction;
- Trigger Systems & Data Acquisition (TDAQ);
- MLOps, Continuous Integration (CI) & CI/CD Pipelines;
- FPGA Design & Programming;
- Hardware Description Languages (HDL) & High-Level Synthesis (HLS);
- Spoken and written English, with a commitment to learn French.
Global Benefits at CERN
Let's get you ready
Be sure to meet the eligibility criteria
- You are a national of a CERN Member State or Associate Member State.
- By the application deadline, you have a master’s degree with 2 to 6 years of professional experience since graduation or a PhD with a maximum of 3 years of professional experience since graduation. You are not eligible with only a bachelor’s degree.
- You have never had a CERN fellow or graduate contract before.
- Please pay attention to the additional criteria and requirements for this specific position and mentioned above.
You will need these documents to complete your application
- Your CV (English or French)
- Any document you consider relevant to your application
- A copy of your most relevant diploma or a certificate of achievement from your school (if you don't yet have your paper diploma)