Education

B.S. in Computer Science

University of California, Santa Barbara
06/2027

Relevant Courses: Foundation Models (graduate), Machine Learning, Generative AI, Operating Systems, Parallel Computing

Experience

NVIDIA

ML Engineer Intern
06/2026 – 09/2026 | Santa Clara, CA
  • •Nemotron post-training data and RL environments

    • ChipAgents | AI for Chip Design

    ML Research Intern
    03/2026 – 06/2026 | Santa Barbara, CA
  • •SoC coverage (formal and functional) with LLM agents
  • •Developed Claude Code features for ChipAgents CLI harness: background subagents, batch mode, AskUserQuestions

    • UCSB NLP Group⁠

    Undergraduate Researcher | Advisor: Xin Eric Wang
    10/2025 – Present | Santa Barbara, CA
  • •Developed a scalable benchmark and RL pipeline for complex, multi-agent robot collaboration and competition
  • •Proposed a grounded task generation method that iteratively creates, simulates, and refines tasks for diversity and difficulty
    • 06/2025 – 09/2025 | San Jose, CA
  • •Made depthwise convolutions 16x faster by optimizing a Rust-based ML compiler for in-memory compute devices
  • •Reduced compile time of tinyML models from 40 mins to 2 secs—then achieved 5.8x faster inference for MobileNetV2
  • •Designed a custom Halide C++ autoscheduler that jointly optimizes tensor fusion, tiling, and memory placement
    • Publications

    Context Rot Bench

    Preprint
    05/2026

    EmToM: An Evolving Benchmark for Embodied Theory of Mind

    Preprint
    03/2026
    Projects & Awards

    FaceTimeOS | Cal Hacks 1st Place Grand Prize⁠

    Won 1st overall at the world's largest hackathon (3k+ particpants)
    Fall 2025
  • •Developed a computer-use AI agent to enable control over personal Macs remotely through FaceTime or iMessage
  • •Built a Flask backend to handle real-time speech interaction through FaceTime's audio
  • •Designed a React + Electron interface to locally prompt the Mac-use AI agent and visualize reasoning steps
  • •Trained a Mixture-of-Experts DQN agent achieving >90% human-level completion rate on Geometry Dash levels
  • •Developed an OpenAI Gym pipeline to evaluate different models and RL algorithms, culminating in a technical report
  • •Built a C++ mod for real-time 60 FPS exchange of screenshots and actions between the AI agent and game
    • Techincal Skills
    Languages — Python, C++, JavaScript, Rust
    Machine Learning — PyTorch, CUDA, ONNX
    Post Training / Inference — VeRL, SGLang, Slime (still learning)