I build AI agents and the backend systems that turn model output into useful work. I bring 15+ years of experience shipping web, mobile, and service products, and now focus on applied AI, retrieval, and workflow automation.
I use AI in software development, research, writing, and translation. My engineering work is in the decisions around it: what tools can run, how evidence is checked, what state survives interruption, and how a result reaches a user.
The projects I write about here are independent work, separate from my employer’s projects. Quantitative research is one application alongside document search, agent platforms, and publishing tools.
START READING
Agent execution and recovery — how I separate reasoning from external calls:
Rust/Rig — typed state transitions, recorded replies, and replay before external I/O:
The series continues with Elixir context, Temporal recovery, shared execution contracts, and preserving state across worker replacement. Each article identifies its implementation scope and distinguishes source review from tests actually run.
BOOKS IN KOREAN
AI와 끝까지 완성하는 Tauri 2 데스크톱 앱
Building, verifying, and shipping an AI-assisted desktop application with Tauri 2, SvelteKit, and Rust. Free chapters:
AI가 만든 자동매매, 검증은 누가 할까?
Data pipelines, backtest accounting, market regimes, and the checks needed to review an AI-assisted trading system. Free chapters:
개발자를 뽑는 개발자
Lessons from reviewing developers’ work and conducting technical interviews. Full text:
Read the books on Folio Books:
PORTFOLIO AND CODE
For engineering collaboration, find me on LinkedIn:
