I recently finished a combined BSc in Computer Science and Biology at UBC, along with a Master of Management. I work on reinforcement learning, mostly around reward engineering and learning under constraints, with side interests in computational biology and AI safety.
A few things I've worked on:
- 🦠 ContagionRL (accepted to TMLR 2026). Agent-based epidemiological models trained with RL to study outbreaks and intervention design.
- 🤖 Hierarchical locomotion in humanoid robots (accepted to AI/CRV 2026). Low-level policies coordinated by a high-level terrain planner.
- 🧬 Protein conformation search (🥈 Silver Medal at iGEM). 80B+ Monte Carlo simulations on UBC's Sockeye cluster to find stable conformations under adversarial conditions.
- 📰 NLP for entrepreneurship. Zero-shot and few-shot classification pipelines at UBC Sauder. Under review.
🔬 Currently a SPAR fellow working on infrabayesian RL for Newcomb-like problems (paper under review). Previously a BlueDot Impact AI Safety fellow.
🌐 More at radman.dev.