Founder · AI Automation Systems — I build pipelines that find demand, produce content, and publish it without a human in the loop.
Remote (Philippines) · Canadian Citizen
Every project below is one stage of the same end-to-end system:
Find demand → Create content → Render media → Ship product → Prove rigor
reddit-lead-gen-analytics — Mines subreddits for recurring pain points and clusters them into a market-intelligence report. Stdlib-only Python, CI, full test suite.
content-mate — End-to-end content pipeline: scrape → create → publish from one CLI, driven by an Airtable control panel. Runs unattended.
viralcut — AI viral-clipping engine: scores footage for hooks, pattern interrupts, and story arcs, then renders vertical clips with animated captions. Strict TypeScript + Remotion, 50+ tests.
fastapi-ffmpeg-stitch — The rendering microservice behind the pipeline: REST API that stitches images + audio into .mp4 via FFmpeg. Tested, CI-backed.
Home-Director — A shipped desktop app: local RAG chatbot over your entire home directory. Electron + Flask + Ollama with a zero-dependency binary vector store. No cloud, no API keys.
riddle-loop — A riddle that finds you while you work. Zero dependencies, cross-platform, 3-OS CI matrix, a test suite where every test encodes a bug that actually happened. The repo to spot-check if you're wondering whether I test my claims.
ai-game-solvers — Autonomous 0-player game solvers, live in-browser: Minesweeper CSP inference, Pac-Man A* with ghost avoidance, Snake BFS + flood-fill autopilot, Word Search 8-directional scanner. ▶ Play live
| Category | Tools |
|---|---|
| Languages | Python, TypeScript, JavaScript, PHP, SQL |
| AI/ML | PyTorch, local LLMs (Ollama), RAG, prompt engineering, LLM output evaluation |
| Media Pipelines | FFmpeg, Remotion, OpenCV, ComfyUI, Whisper |
| Backend | FastAPI, Flask, REST APIs, MySQL, SQLite, Airtable |
| Algorithms | A* pathfinding, CSP, BFS, flood fill, heuristic search |
| Dev Tools | Claude Code CLI, Cursor, Git, GitHub Actions |
- Tear it apart — I don't just use tools. I break them, audit their outputs, and find where they fail.
- Build the pipeline — If a process is manual and repetitive, I script it. Python + CLI tools + automation.
- Evaluate relentlessly — Every AI output gets audited for logical failures, prompt drift, and edge cases before it touches production.
- Document everything — If it isn't written down, it doesn't exist.
- Autonomous content pipelines (demand discovery → creation → publishing)
- Multi-agent AI orchestration patterns
- Local LLM deployment and evaluation
📍 Remote · 🇨🇦 Canadian · 📧 Clawd2Bot@Gmail.com
"There's got to be a better way." — My approach to every problem.