What I do
Latency Sprint
Cut inference latency without accuracy loss using profiling, quantization, and TensorRT/ONNX.
Edge Deployment
Ship models to Jetson/Orin/CPU-only edge with monitoring, rollbacks, and CI/CD.
Advisory & Reviews
Architecture reviews, eval harnesses, and hands-on guidance for CV/Gen-vision teams.
Featured case studies
Latest writing
Worked with
Automotive OEM
Robotics Startup
Retail AI
Gen-Vision Lab