Limitations in Planning Ability in AlphaZero
NeurIPS 2024Daisy Xinlei Lin, Brenden Lake, Wei Ji Ma
Paper →
AI Researcher @ Amazon AGI SF Lab
I build human-like AI agents, because the only thing more fun than being human is trying to recreate parts of it in silicon. I got my PhD at New York University, where I studied planning and reinforcement learning: how minds and machines decide what to do next when the world is messy, uncertain, and adversarial. These days I focus on LLM post-training, multimodal reasoning, and computer-use agents -- the kind of systems that don't just talk, but act.
Major contributor to Nova Act — an AWS service for building reliable AI agents that automate UI workflows with 90%+ task reliability. Trained the model using RL in synthetic web environments.
Daisy Xinlei Lin, Brenden Lake, Wei Ji Ma
Paper →Zheyang Sam Zheng*, Xinlei Daisy Lin*, Jake Topping*, Wei Ji Ma
Paper →Jordan Lei, Jeroen Olieslagers, Nastaran Arfaei, Xinlei Lin, Wei Ji Ma
Paper →When I'm not teaching models what to do, I'm usually snowboarding, hiking, baking desserts, or being supervised by cats who believe they are my bosses.
I moved from NYC to SF to join the RL research team at AGI SF Lab.