A collection of project demos, research visualizations, and AI system architectures from my work at AGI Inc and Stanford.
Hackathon-winning AI system controlling macOS through MCP interfaces
Benchmark results showing 76% reliability improvement using GRPO training
Scalable web agent framework architecture handling 12K+ downloads
Real-time visualization of policy distillation in reinforcement learning
Live demonstration of AI agent navigating complex web interfaces
Published research paper on comprehensive web agent evaluation framework
NVIDIA H200 cluster setup for large-scale RL training with Ray
Real-time monitoring dashboard for reinforcement learning experiments