Hello World!
I'm Hangyu Feng
> Deep Learning Software Engineer
> Graduate Researcher @ Ai2
> M.S. Applied & Computational Mathematics, UW
3+ years at Intel building model deployment tooling, inference benchmarking pipelines, and automated CI/build systems for deep learning workloads; specialized in model evaluation and optimization across PyTorch, TensorFlow, and OpenVINO.
Recently completed an M.S. in Applied & Computational Mathematics at the University of Washington, with full model lifecycle experience — teleoperation data collection, distributed fine-tuning, and on-device inference of VLA foundation models on physical robot hardware at the Allen Institute for AI (Ai2).
Fine-tuned a vision-language-action model to control a robot arm across simulation and real hardware, and curated the teleoperation datasets it trained on.
Benchmarked and optimized neural-network inference on Intel's FPGA accelerators, and built the C++ deployment tooling and CI/reporting pipelines around it.
Built full-stack features, test suites, and anti-fraud systems for two fintech products, integrating them through a shared API.