Gergely Racz

I design and ship production AI systems — not prototypes. PhD in Engineering from the University of Cambridge, 10+ years in AI, starting in machine learning for automated driving and now building the systems companies actually run their operations on.

Recent work: two production RAG systems for a $3B/yr multinational's engineering team (multimodal retrieval over 5,000+ technical documents — in daily use), autonomous AI agent systems that run scheduled and monitored in production, and end-to-end sales automation pipelines (scraping → AI scoring → CRM → outreach).

What I build:

- RAG & knowledge systems — retrieval over your documents with source citations, low hallucination, enterprise-grade (pgvector, rerankers, multimodal)

- AI agents & agentic workflows — LangGraph, MCP, multi-agent orchestration; reliability-first: evals, error handling, monitoring

- Full-stack AI products — FastAPI, Next.js, Postgres/Supabase; agency-scope delivery solo, in weeks not months

- Business automation — n8n, Make, Airtable; lead generation, voice agents, LinkedIn/email outreach systems.

I work as a hands-on technical partner: I translate between business goals and architecture, propose automations you didn't ask for, and ship things that survive contact with production. Documented builds on my YouTube channel (AI Lead Flow Automations)

Based in Hungary (CET) — full overlap with EU business hours.