I work on the data-processing infrastructure behind the LHCb experiment at CERN.
I'm a software engineer at CERN's LHCb Computing Section, working on the systems that transform petabytes of raw LHC collision data into datasets that physicists use to make discoveries and observations about our Universe. I started out as a physicist myself: my PhD produced the observation of a new charmed baryon, and the tooling and central infrastructure I wished I'd had during that time became the reality of what I do now. These days that means a few things, including but not limited to: Analysis Productions, the hugely successful platform behind LHCb's Run 3 dataset production, and CWL-based workflows for DiracX, the next generation of the DIRAC grid middleware.
GitHub · CERN GitLab · LinkedIn
Selected work
- LHCb Analysis ProductionsSelf-service data processing for hundreds of LHCb physicists; the official route for LHCb Run 3 data.
- Autonomous operations serviceFrom single-user log-analysis script to concurrency-safe service with guarded, audited interventions.
- LbAgents: Automatic Agentic Workflows in LHCbExtensible, pluggable, and testable proof-of-concept agentic workflows using LangGraph that diagnose CI failures and stuck productions, with guardrails: every write action is gated on human review. Plugs into infrastructure and highly configurable with existing MCP or user-specified tools.
- DiracX / CWL integrationOpen, portable workflow descriptions for the next generation of the DIRAC grid middleware.