Research

Machine learning applied to physical and engineered systems — reinforcement learning, physics-informed models, and optimization.

Interests

Projects

ARAI — Reinforcement Learning for Prosthetic Hand Control

Investigating the application of reinforcement learning and virtual reality to prosthetic hand control for sub-radial amputees. Uses Unity ML-Agents to train adaptive actuation policies in augmented reality environments, with the goal of improving prosthetic responsiveness and user outcomes. This work contributed to a peer-reviewed publication.

Reinforcement learning · Unity ML-Agents · AR/VR — source

Physics-Informed Neural Networks for Jet Engine Compressors

Applying physics-informed neural networks to improve the thermodynamic efficiency of jet engine compressors. Embedding governing physical laws directly into the network architecture yields models that are both data-efficient and physically consistent, reducing reliance on costly simulations.

PINNs · Thermodynamics · Aerospace — source

Adaptive PageRank Optimization

PageRank ranks nodes in a graph using a fixed damping parameter applied uniformly across all nodes. This project asks whether giving each node its own locally adaptive damping factor can improve convergence and ranking quality, framed as a constrained optimization problem over the space of stochastic matrices — drawing on Perron–Frobenius theory, Markov chain perturbation analysis, and fixed-point arguments.

Graph theory · Optimization · Linear algebra — source

Elsewhere

Unreasonable Precision

A YouTube channel documenting engineering projects start to finish — the thought process, the dead ends, the technical work, and the result. No polished highlight reels, just the full solve.