About me

Hi, my name is Marek Kucak. I design production-grade Generative AI systems — secure, cost-controlled, and ready for enterprise scale.

My focus is turning prototypes into real systems that are governed, observable, and financially sustainable.

I’m currently an AI Architect / Data Scientist at Siemens Healthineers, where I build GenAI, RAG, and data engineering solutions on Databricks, with an emphasis on security, evaluation, and FinOps.

What I do

• Architect secure RAG systems using Unity Catalog, Vector Search, and Delta.

• Build agentic workflows with Genie, LangGraph, and structured tool execution.

• Implement FinOps into CI/CD — cost-gates, model arbitration (DSPy), and token optimization.

• Deploy ML and GenAI systems in complex environments (cloud, hybrid, or restricted networks).

• Build reproducible pipelines using Python, Spark, SQL, Rust, Linux, Docker/K8s.

My background

• Over a decade across healthcare, banking, and insurance (Siemens Healthineers, Erste Group, Swiss Re).

• PhD in Crisis Management & Econometrics — trained to think in risk, constraints, and system robustness.

• Coding since age 10 (programming + math program).

• Linux enthusiast and builder of practical engineering workflows.

What I’m working on today

• Secure enterprise RAG architectures

• Databricks Agentic Systems with evaluation-first design

• AI FinOps frameworks for CFO-approved deployments

• Educational content & practical engineering tutorials