I’m a student engineer focused on turning AI/ML ideas into usable software systems — from model experiments and local-first tooling to mobile/web products that can actually be tested, shipped, and improved.
AI product engineering · full-stack systems · developer tooling · local-first software · applied ML
- AI products: model-backed workflows, intelligent automation, evaluation, and human-in-the-loop product design.
- Software systems: end-to-end mobile/web/backend implementation with a production mindset rather than demo-only prototypes.
- Applied ML & edge computing: computer vision, embedded devices, and practical inference under real hardware constraints.
| Project | What it solves | Core stack |
|---|---|---|
| DeskGraph | A privacy-first, local computer context graph that lets AI agents search narrowly authorized local context without uploading file data by default. | Rust · Tauri · SQLite · local-first architecture |
| Google Play Preflight Skills | A policy-aware preflight skill for AI coding agents that catches likely Android / Expo / React Native / Flutter review blockers before Play Store submission. | Agent tooling · Android policy · Expo / React Native |
| Smart Microwave Oven | Contactless microwave input using computer vision, Raspberry Pi hardware control, and on-device ML-oriented inference. | Python · TensorFlow / Keras · OpenCV · Raspberry Pi |
Automatically refreshed from GitHub API data. When private-repository access is configured, this card aggregates public + private default-branch commits while never rendering private repository names or commit messages.
Build → measure → ship → learn.