Mechanical Engineer specializing in ML and AI, bridging physical-systems engineering with machine learning.

I'm as comfortable modeling a thermal or dynamic system as I am training a reinforcement learning policy or building an autonomous agent. I like systems that operate on their own: agents that research around the clock, and RL policies that learn to act under uncertainty.

• autonomous-research-agent. A self-learning AI that crawls knowledge, runs research loops, writes structured memos, and fine-tunes itself weekly. Qdrant vector memory, RAG, daily self-assessment.
• RL trading research (private). PPO, RecurrentPPO (LSTM), and a JAX PureRL pipeline on a single RTX 4090 at 1.1M transitions/sec.
• Circadian (private). A SwiftUI iOS app for circadian-rhythm tracking, with HealthKit integration, Live Activities, and adaptive scheduling.

ML / AI: Python · PyTorch · JAX / Flax · Stable-Baselines3 · Qdrant · CUDA Engineering: physical-systems modeling · thermodynamics · numerical methods · MATLAB Also: SwiftUI

LinkedIn

_{Much of my strongest work lives in private repos. Happy to walk through any of it on request.}