Event Reconstruction Physicist (EP-FCC-2026-166-GRAP)
Core
Develop a detector-independent graph data model and navigation API for high-energy physics event reconstruction, integrating graph algorithms and machine learning for clustering and particle identification.
Role type
Senior IC experimental physicist (high-energy physics software)
Builds
A portable, experiment-independent library for event reconstruction used in CMS and FCC software stacks
Domain
High-energy physics, computational physics, machine learning
Deliverable
production ML models
Required skills
C++20, event reconstruction, detector simulation, particle flow, heterogeneous computing (CUDA/SYCL/Alpaka), graph algorithms, machine learning (PyTorch), Python
Preferred skills
Experience with CMSSW, Key4hep, Gaudi, transformers, graph neural networks
Technologies
C++20, PyTorch, CUDA, SYCL, Alpaka, CMSSW, Key4hep, Gaudi, EDM4hep
Responsibilities
Generalize the CMS MC-truth graph into a detector-independent library; Develop bindings for CMS and FCC stacks; Extend the data model with sparse association and pileup handling; Develop reconstruction-to-truth association algorithms; Define reconstructible objects from truth graphs; Apply graph algorithms and ML for clustering and particle identification; Build training datasets and validate models; Contribute code to official repositories with tests and documentation; Present results in meetings and publications; Provide team supervision.
Seniority
Senior, hands-on IC with team supervision
