My current focus is AI engineering — specifically LLM integration and AI-powered products. After shipping the Gemini integration in Iscanow, I got hooked on the possibilities.
Right now I'm exploring AI agents and autonomous workflows: systems that don't just answer questions but take actions, call tools, and complete multi-step tasks. The design questions are fascinating — how do you make an agent reliable? How do you give it just enough autonomy?
I'm also continuing to sharpen my core stack: Next.js, TypeScript, and Node.js remain home base, with blockchain as a constant source of interesting problems.
Wherever this leads, the approach stays the same: pick a problem that feels slightly too hard, build something real, and learn everything the process teaches me. More updates soon.