> about
University of Pennsylvania M&T Program '27
> Grew up between Boston and London. Now based in Philly.
> I build AI systems, do ML research, and invest in startups.
> Research Engineer @ Labelbox — RL, evals, post-training.
> Contributed to models at Anthropic, DeepMind, and Meta FAIR.
> Multi-agent RL research @ Penn ML Theory Group.
> Venture Partner @ Contrary Capital.
> Chess, poker, cars, food, impressionist art.
> experience
> research
Research on multi-agent reinforcement learning systems, exploring coordination, competition, and emergent behavior in complex multi-agent environments.
Developing vision-language model applications for industrial robotics, bridging perception and language understanding for robotic manipulation and task planning.
> projects
High-performance packet filtering system leveraging NVIDIA CUDA for GPU-accelerated network traffic analysis, achieving significant throughput improvements over CPU-bound BPF implementations.
Real-time vehicular crash detection system using distributed IoT sensor networks, accelerometer fusion, and edge computing for instant emergency response triggering.
> blog
Lessons from building evaluations for frontier AI models — why quality means making questions real, not harder.
Mapping three value-creation paths to founder archetypes for evaluating early-stage startups.
On finding direction — from high school confusion to ML research and early-stage investing.