I am Yohan Chatelain, a scientific associate at the Krembil Centre for Neuroinformatics, Centre for Addiction and Mental Health (CAMH), since March 2026. I am based in Montréal, Canada, and work with the Epics lab.

Research and current work

My research connects high-performance computing, scientific software, and numerical reliability. I study LLVM instrumentation, floating-point and stochastic arithmetic, mixed and reduced precision, and reproducibility in AI and neuroimaging pipelines.

I work on the Brain Health Data Challenge platform, developing computing infrastructure and evaluation methods for responsible, collaborative research with sensitive health data. My work also examines the numerical stability of neuroimaging models and longitudinal analyses.

I created PRISM and developed PyTracer. I also contribute to Verificarlo and its instrumentation tools, and to Fuzzy PyTorch. My publication list and CV provide the corresponding research record.

Previous appointments and education

I was a postdoctoral researcher at CAMH from January 2025 to February 2026, and at Concordia University from September 2020 to December 2024 in Big Data Infrastructures for Neuroinformatics.

I completed a PhD in Computer Science at Université Paris-Saclay / UVSQ in 2019, a Master’s in High-Performance Computing and Numerical Simulation in 2016, and a Bachelor’s in Computer Science in 2014. My doctoral work at LI-PaRAD investigated numerical debugging and reduced-precision optimization of scientific applications.

Awards and research support

My research record includes the ACM REP 2024 Best Paper Award, the Concordia Horizon Postdoctoral Fellowship (2020–2022), and an AccelNet IN-BIC research grant (2021). Details and associated work are listed in the CV.

Profiles and collaboration

I use ORCID 0000-0001-7023-250X for researcher identification. See Google Scholar, GitHub, and LinkedIn for publications and software.

Contact me about numerical-analysis tooling, scientific software, and research collaboration.