Vision models on edge devices

Professional work / Black Sesame Technologies

The challenge: preserve model behavior when moving from training to a camera runtime.

My contribution

Connected data preparation, training, quantization, export validation, native integration, device testing, and model replacement checks across segmentation, image matting, monocular depth, detection, and camera-oriented image processing.

Design decisions

Model architecture, I/O contracts, quantization behavior, runtime assumptions, and device debugging were treated as one design space rather than separate team boundaries.

Representative stack

Python, PyTorch, TFLite, ONNX, TensorRT/ONNXRuntime, OpenCV, C++, Android NDK, CMake, Docker, and device-side debugging.

Portrait imaging & computational photography

Professional work / Black Sesame Technologies

The challenge: combine depth, subject boundaries and rendering into a controllable portrait effect.

My contribution

Worked across depth-aware bokeh, portrait segmentation and matting, subject protection, optical highlights, GPU rendering, and real-time composition using detection, optical-flow tracking, camera guidance, and application state behavior.

Design decisions

Final image quality depends on the interaction between model output, rendering parameters, long-tail boundaries, motion, device constraints, and review criteria. The work therefore included visual debugging, hard-case analysis, batch comparison, and release-quality review artifacts.

Representative stack

Monocular depth, image matting, detection and tracking, OpenGL/GLSL, C++ rendering, alpha compositing, OpenCV, TensorRT, ONNXRuntime, and Android camera workflows.

Independent work: agent & multimodal workflows

Public reference code and browser-local applications. The workflow uses scripted providers, and the photography app uses deterministic analysis; neither demonstrates a live LLM deployment.

Public evidence

Implemented a provider-neutral Python workflow for retry, fallback, structured-output validation, confidence gates, deterministic traces, and hard-case evaluation. Built a browser-local Photography Mentor for image diagnostics, retrieval, practice planning, and review state.

Design decisions

Routing and generation are not enough. Expected failures need explicit policy; invalid outputs must be rejected; low-confidence results require review; and every decision path should leave a trace that can be reproduced in a hard-case suite.

Representative stack

Python, dataclasses, provider protocols, validated output contracts, unittest, local retrieval, deterministic image analysis, and browser-local persistence.

Product system: Business Learning Studio

A privacy-first, offline-capable workspace that turns ambiguous business questions into evidence, experiments, and decisions.

Public evidence

Built a complete browser-local learning loop with diagnostic planning, four practice modes, transparent feedback, mastery tracking, 32 original knowledge cards, and a five-gate decision coach.

Design decisions

The product treats learning as observable performance rather than content completion. User state remains in the browser, runtime dependencies are vendored, and release automation rejects screenshots, local paths, private-source markers, and non-allowlisted files.

Representative stack

Progressive Web App, service-worker offline cache, local storage, deterministic coaching rules, generated JSON knowledge artifacts, Node.js contract tests, privacy audit, and GitHub Actions.

Evaluation & handoff

Make the release decision reproducible for the next engineer.

Translate ambiguous product or field issues into a technical scope, failure taxonomy, hard-case set, prototype, measurable acceptance criteria, regression path, and release decision. Preserve reproducible traces and turn working patterns into reusable tools, documentation, and handoff material.