jigonyoo
Remote: Yes - async, in writing (no calls)
Willing to relocate: No
Technologies: Python, SQL, Postgres/Supabase, BigQuery, dbt, DuckDB, web scraping/extraction, LLM & agent evaluation (RL environments, graders), agent tool-call safety (idempotency, approval gates, audit logs), Korean-language data/OCR
Résumé/CV: https://jigonyoo.com (public MIT repos with their measured results)
Email: [email protected]
I build data and AI checks you can verify: quality gates that stop a bad batch before it reaches a dashboard, and evaluations that measure what an agent actually did, not what it said. Recent: an RL environment on Prime Intellect's Environments Hub that measures how often models pay the same refund twice - 14 vs 3 vs 3 duplicates across 96 rollouts each, reproducible from a fresh clone. Freelance/contract, fixed scope preferred.