research software engineer — I make ML work in prod, not just in notebooks.
currently building research platform infrastructure for biological data visualisation @ Oxford's Centre for Human Genetics, on the MDV platform. previously built MLOps + geospatial pipelines for epidemiological forecasting at LSHTM, worked on protein language models in biotech, and did a stint at a YC startup.
i like the space between research and production. most of my work involves turning messy scientific workflows into something deployable.
- ChatMDV — an LLM-powered interface for exploring biological data, with a focus on provenance and sandboxed execution
- a distributed jobs framework for MDV with a pluggable executor abstraction (Local + Slurm backends), so analysis jobs can run on HPC as easily as locally
- vLLM / local LLM inference across HPC and OpenStack environments
mdvtools— packaging and shipping it on PyPI
