AI builder.
Data engineer.
Building in public.
Chris S. Yoon. 10 years across QA, SDET / automation, and Data Engineering — now using AI tools to build products I actually want to use.
Background
10+ years
QA → SDET / automation → Data Engineering at theScore / ESPN Bet.
Building with
AI tools
Claude Code, Codex, and the full AI-assisted dev loop — learning in public.
Current build
opb-poker
A poker performance backend I built because the right tool didn't exist.
The problem
The tools I needed
didn't exist.
So I built one.
AI tools
Session reviews forget everything between sessions. Next time at the table — same mistakes, no memory, no growth.
Solvers
Expensive, complex, and built around optimal ranges — not around understanding your own tendencies and blind spots.
Friends
Asking for advice means sharing your edge with someone you'll play against. Information stays inside.
“I play seriously. I want to improve. I needed something that actually remembers me.”
opb-poker
A backend that knows your patterns.
Three-layer data architecture: raw session packets ingested and deduplicated, normalized into canonical truth, derived interpretation built on top — each layer separate, inspectable, and independently testable.
Entity-centric PostgreSQL schema — player, session, tournament, hand, pattern, and intervention as first-class entities. Cumulative state across sessions, not stateless one-off review.
TDD-first: output contracts defined before implementation. Pytest regression suite covers ingestion through runtime serving. Playwright E2E on the consumer surface. Operator QA review gates before anything ships.
“I started wanting to build for everyone. Then I realized: I am the audience. If I can’t satisfy myself first, I can’t satisfy anyone like me.”
Work
Active build
Private AI Poker Backend
Three-layer data pipeline (raw → normalized → derived) over an entity-centric PostgreSQL schema. TDD-first build with deterministic output contracts, Pytest regression suites, and operator QA gates.
Ingestion pipeline treats each upload as a session packet: parsed, deduplicated, and reconciled into canonical truth before any derived layer runs — provenance preserved end-to-end.
Entity-centric PostgreSQL schema: player, session, tournament, hand, pattern, and intervention as first-class entities. Raw, normalized, and derived layers kept separate and independently queryable.
TDD-first across the full stack: output contracts defined before implementation, Pytest regression suite from ingestion through runtime serving, Playwright E2E on the consumer surface.
Past work
About
Chris S. Yoon
10 years across QA, SDET / automation, and Data Engineering — most recently Senior Data Engineer at theScore / ESPN Bet. Now building AI products: current main build is opb-poker, a backend-first performance system built on real session ingestion, cumulative memory, and deterministic AI-driven outputs.
Recruiter path
Career context, then current work.
Background first, then current work, then LinkedIn or GitHub if you want more.