I am a Member of Technical Staff at Microsoft AI, working on health-related projects, evaluations, pretraining and midtraining.
Previously I was at EIT, developing foundation models for science and healthcare. Before that I was at Meta and DeepMind developing RL training algorithms and environments for code modelling and generation. And before that I worked at Jukedeck building some of the earliest ML models for symbolic music generation.
CWM: an open-weights llm for research on code generation with world models
Nash learning from human feedback
Faster sorting algorithms discovered using deep reinforcement learning
StructureNet: inducing structure in generated melodies