Skin-cancer detection with a verification step.
Skin-cancer detection with a separate verification layer. I trained the vision model and built the system that lets a doctor inspect its attention maps.
- PyTorch
- FastAPI
- Grad-CAM
I build medical AI, agent systems, and developer tools. Before that, I spent nine years leading teams and delivering products for international clients.
Explore selected workSkin-cancer detection with a separate verification layer. I trained the vision model and built the system that lets a doctor inspect its attention maps.
View project9 years leading development
15+ international clients
~10 core team members
A closer look at the problems, the systems, and my contribution.
Skin-cancer detection with a separate verification layer. I trained the vision model and built the system that lets a doctor inspect its attention maps.
Local agent loops with a planner, fresh workers, and a critic. A file-based rubric keeps each cycle accountable, with a dashboard to follow the work.
A production pipeline that researches, writes, and publishes AI news across two live sites. Three white-label deployments sold.
Deep learning for basal cell carcinoma histopathology and real-time melanoma/BCC detection for Mohs surgery. Three-tier classification beyond binary labels; computer-vision models, two patents pending.
Explore projectMulti-agent AI platform that turns prompts into full-stack app prototypes. Live previews, automated evaluation loops, and model-agnostic orchestration across Claude, GPT, Gemini, and local models.
Explore projectTerminal statusline for Claude Code. Real-time quota tracking, zero dependencies. Built for devs who live in the CLI.
Explore projectMulti-AI group chat for consensus. Multiple models discuss topics together and converge through structured rounds.
Explore projectGitHubGamified Python learning platform. 30 progressive levels, real code execution in-browser via Pyodide, AI-powered evaluation. Zero setup, completely free.
Explore projectGitHub
Edward Yi / Dallas-Fort WorthI spent 9 years in blockchain development, most recently leading client delivery at a dev firm: a core team of ~10, coordinating with an overseas office to ship products for 15+ international clients across DeFi, wallets, and NFT marketplaces.
I now build AI products full-time. Multi-agent platforms, developer CLI tools, medical AI, and full-stack applications.
As a kid, I set up and hosted three 32-player Counter-Strike servers: vanilla, Superhero mod, and Warcraft 3 mod. They were consistently full, and I sold admin and level passes.
More about my backgroundBuilding products, leading people, and staying close to the work.
Building AI products full-time: medical AI for skin-cancer detection, developer tools, and multi-agent systems. Research selected for the AAD Innovation Academy 2026 (Digital/Electronic Technology) for three-tier deep learning on basal cell carcinoma histopathology. Hackathon builder at Anthropic's Built with Claude and OpenAI's Build Week.
Led the South Korea branch. Core team of ~10, coordinating delivery with a Dubai office for 15+ international clients. Built DeFi protocols, crypto wallets, NFT marketplaces, and data-streaming infrastructure. Now a technical contact / referral partner.
Founded and operated LED lighting solutions company in Austin, TX. Hardware and software product development, supply chain management, and B2B sales.
For engineering roles, collaborations, or a conversation about what you’re building.