Machine learning for medical imaging and robotics.

I worked on cardiovascular imaging: building ML models for echocardiograms, and data + inferencing pipelines. A large part of the job was getting those models to run reliably across a hospital system's whole imaging archive, which meant rebuilding inference around multiprocessing and multiple GPUs.

I built the perception stack for a forestry robot on an AgileX SCOUT platform, the part that decides which ground is safe to drive over, so crews can clear wood waste before it becomes fuel. I fine-tuned models and optimized inference on an NVIDIA Orin AGX, which is what made it fast enough to use in the field.

I wrote validation pipelines for Direct Debit and Credit Transfer channels against the FPS and HKICL specs, integrated and built Spring REST clients with the BPay API to speed up bill payments, and led three SDETs shipping a Spring Boot test pipeline across three testing and production deployment regions.