Portrait imaging & computational photography
Connect segmentation, matting, monocular depth and GPU rendering. Investigate boundary artifacts, subject protection and visual quality in difficult scenes.
PyTorch / OpenCV / C++ / OpenGL & GLSL
Computer vision & applied AI
Computer vision.
Built for real devices.
I build camera AI at Black Sesame Technologies, connecting portrait matting, depth estimation and rendering to C++/Android deployment and release evaluation.
I'm interested in Applied AI and Forward Deployed Engineering roles, including multimodal systems where model quality, product requirements and runtime constraints meet.
Professional experience
Camera AI, end to end
Vision models on edge devices
Take models through quantization, export validation and native camera integration, with device debugging and regression checks before replacement.
TFLite / ONNX / Android NDK / CMake
Employer work is summarized without customer identities, proprietary code or internal performance data.
Independent projects
Code you can inspect
01 routing primary selected 02 generation primary started 03 generation primary exhausted 04 routing fallback selected 05 generation fallback started 06 validation fallback passed 07 review_gate fallback passed 08 completion fallback accepted
Python reference
Reliable Agent Workflow
Retries, fallback, output validation and review gates with reproducible failure-path tests. Scripted providers; no live LLM integration.
Browser application
Photography Mentor
Local image diagnostics, source-linked retrieval and practice tracking. Deterministic image analysis and coaching rules; no model API required.
Also built: Business Learning Studio, an offline decision-practice workspace.