I build AI systems that have to work—not just in demos, but in production where mistakes cost money and trust.
What I Do
Currently Founding AI Engineer at Parcha, where I’m building AI agents for compliance automation. Most recently shipped Grep.ai—an AI-powered due diligence platform built on Anthropic’s Agent SDK that researches businesses across 190+ jurisdictions in under 3 minutes.
The hard part isn’t getting LLMs to work. It’s making them reliable enough to bet your business on.
My focus areas:
- Production Agent Systems — Handling real-world complexity beyond the “happy path”
- Evaluation Frameworks — Designing rigorous tests for high-stakes AI (compliance, finance)
- Reliability — Reducing false positives by 90%+ without sacrificing recall
Background
15+ years in data systems across fintech, automotive, and enterprise.
Previously:
- ML Team Lead at Carvana — Led machine learning team, shipped models serving millions of users
- Senior MLE at Augment CXM — 5 years building core algorithms and data science platform
I specialize in the messy intersection of research-quality AI and production-grade engineering.
Get in Touch
- GitHub: KDmytro
- LinkedIn: kdmytro
- X: @kdmytro
- Email: kdmytro@gmail.com
Based in San Francisco, CA.