STEFAN
I design, ship, and operate AI systems end-to-end.
I translate ambiguity into intelligent, production-grade systems that learn fast, scale cleanly, and ship reliably.
When the path isn't obvious, I architect it. When systems get messy, I make them elegant. When it must work, I make it robust.
- 6 years of experience building reliable systems
- Since age 7 I have lived on computers
- 40+ applications running in production
Stefan - Full-Stack AI Systems Engineer
I started in assembly and C++, moved through Java, and that low-level curiosity still drives how I architect AI systems today. These days I live in Python, TypeScript, and cloud-native stacks to design, ship, and steward ML-powered products.
My focus is AI infrastructure: retrieval pipelines, LLM orchestration, observability, and the services that keep them online.
I care about deterministic releases, human-in-the-loop safeguards, and closing the gap between prototypes and production.
Tech Stack
AIAI & LLM Systems
Graph-grounded reasoning, agentic pipelines, and production LLM orchestration.
OPSMLOps & Pipelines
Governed training + deployment loops, observability, and reproducible data workflows.
BEBackend & APIs
Event-driven services and realtime APIs that keep AI features production-ready.
PLTPlatforms & DevOps
Delivery infrastructure, telemetry, and automation for enterprise rollouts.
DBDatastores & Vectors
Operational stores and vector memory built for retrieval-heavy systems.
UIProduct & Interfaces
Admin control planes, analytics dashboards, and polished prototypes.
Project archive
Client work lives off the grid
Every product I ship is built under NDAs for venture studios, so the real builds stay private. If you want a peek at the experiments, automations, and AI tooling I tinker with after hours, my GitHub is always open.
Education
Academic foundation
Bachelor's Degree in Information Technology
Engineer of Information Technology.