How I Work
AI-Driven Development
I architect solutions and pair with AI under a strict system of standards. Not "copy from ChatGPT" — an engineered process with automated quality gates.
Pipeline
Code
Codex standards auto-loaded per stack
Review
Self-review + code review on every change
Push
Conventional commits, feature branches
CI
Lint + typecheck + tests — must be green
Deploy
Auto-deploy to production
Codex
A system of code quality rules that both I and AI follow. Every project, every commit.
READ before WRITE
Read the file you're changing + 2-3 similar files in the project. Understand patterns before writing.
VERIFY, don't guess
Check library docs for every API call. Never guess method signatures or behavior.
ONE thing at a time
Finish one task completely before starting the next. Minimal diffs, focused changes.
CHECK after writing
Re-read your own diff. Verify imports exist, types match, edge cases are handled.
Analogy
Manual Counting
In development: writing every line by hand, catching bugs with your eyes.
Slow, error-prone, but full control.
Calculator
In development: ChatGPT / Copilot — fast answers, but no verification. Hallucinated APIs go to production.
Faster, but you trust blindly. Wrong formula? You won't know until it breaks.
1C Accounting → AI-Driven Dev
In development: standards, automated reviews, CI/CD. I make decisions, the system enforces quality.
1C didn't replace accountants — it made them 10× more productive. Same here.
Want to see it in action? Let's talk.
Message on Telegram