I build reinforcement-learning controllers, physics-based simulation environments, and sim-to-real workflows for physical systems. My work focuses on continuous-control reinforcement learning, policy regularization, robotics, adaptive optics, and reliable hardware-aware control.
During my doctoral and postdoctoral research at the University of Ottawa, I developed State-Adaptive Proportional Policy Smoothing (SAPPS) and evaluated learning-based control methods across MuJoCo benchmarks, wavefront-sensorless adaptive optics, and a real Crazyflie nano-quadcopter.
I am open to full-time roles in robotics, reinforcement learning, simulation, and applied machine learning for physical systems. I am authorized to work in Canada and open to relocation across Ontario.
A policy regularization method designed to reduce unnecessary action fluctuations while preserving responsiveness as system states change rapidly.
- Evaluated across MuJoCo continuous-control tasks, adaptive optics, and real-world quadcopter control
- Implemented with PPO and compared with standard PPO and policy-smoothing baselines
- TechRxiv preprint; manuscript currently under revision
An open-source, Gymnasium-compatible environment for reinforcement-learning control of wavefront-sensorless adaptive optics in optical satellite communication.
- Configurable atmospheric dynamics, observations, actions, rewards, and episode structures
- Optical propagation, deformable-mirror control, focal-plane sensing, and single-mode-fiber coupling
- PPO, SAC, and DDPG implemented from scratch in PyTorch
An end-to-end simulation-to-hardware PPO workflow for Crazyflie 2.1 altitude control.
- Multi-seed simulation training followed by 100 real-hardware episodes
- Real-time telemetry, safety constraints, automatic landing, and hardware integration
- Built with Python, PyTorch, cflib, Crazyradio, and the Flow Deck
- Reinforcement learning: PPO, SAC, DDPG, actor-critic methods, continuous control, policy regularization
- Simulation: Gymnasium, MuJoCo, custom physics-based environments
- Machine learning: Python, PyTorch, Tianshou, NumPy, pandas, Matplotlib
- Research computing: Linux, Bash, HPC, SLURM, Weights & Biases
- Robotics: Sim-to-real evaluation, Crazyflie, hardware integration, robotic skill learning
- Action-Regularized Reinforcement Learning for Adaptive Optics in Optical Satellite Communication — Journal of the Optical Society of America B, 2026
- Adaptive Policy Regularization for Smooth Control in Reinforcement Learning — TechRxiv preprint, 2026
- Reinforcement Learning Environment for Wavefront Sensorless Adaptive Optics in Single-Mode Fiber Coupled Optical Satellite Communications Downlinks — Photonics, 2023