Advisory · Build · Enablement
Data platforms and LLM systems that hold up in production.
Independent data and AI architect. From raw data to a platform that holds up in production — then advisory as your team takes it forward.
Focus areas
Data platforms
From raw, unused data to a queryable, governed foundation — end-to-end, on the stack that fits your scale and budget.
LLM & agent systems
From prototype to a production agent you can trust — retrieval, evals, and multi-agent architecture that don't silently fail.
Data foundations for AI
From an AI that guesses to one that answers correctly — data modeling, quality, and the knowledge layer underneath.
Case studies
Real engagements in software architecture, data engineering, and AI.
002 / 2025 A bigger model wasn't the answer: what actually makes an analytics agent hold up A lean prototype measured that for cents per analysis and kept the budget out of the wrong investment.
Data Agents PoC
→ 001 / 2025 A serverless AWS data platform that turns multi-TB IoT data into analytics — for the price of a team lunch Designed and built a lean, serverless data platform that lets a smart-buildings IoT company store and analyze terabytes of sensor data at a near-trivial running cost — and gave their team a developer workflow that spins up new data projects in minutes. Data AWS
→ Thoughts
Thinking out loud about tech, data, and the consulting life.
The biggest model is rarely the right one Why small, specialized models automate most business processes better than the most expensive frontier model, and what it takes to get there.
machine-learning technology strategy
5 min What Are We Actually Trying to Achieve? Why chasing the next technology hype is a trap — and why the right technology decision starts with the question nobody likes to ask. technology strategy consulting
2 min How a Severed Fiber Optic Cable Sparked My Passion for Data How analyzing faults at Deutsche Telekom in Bonn — and a perceptron trained in 2015 — turned into a lasting passion for turning data into real value. data machine-learning career
6 min Hello World A long-overdue first post on why the tech guy who builds for everyone else finally built something for himself. introduction consulting behind-the-scenes
3 min Services
Three ways to work with your team — advise, build, enable.
01
Advisory
Helping you understand what to do.
- Sparring partner for engineers and managers.
- Technology evaluations — a detailed report with comparison, pros and cons, and a clear recommendation.
- Reviews and audits of data & AI architecture and applications.
02
Build
Building the thing — on real data, in production.
- A proof of concept, concept, or design you can build on.
- End-to-end data & AI platforms.
- Custom web, data, and AI applications.
03
Enablement
Equipping your team to carry it forward.
- Technology training so staff use new tools right from the start.
- Ideation workshops to find new solutions to old problems.
- Strategy workshops to build a roadmap for the organization.
Explore specific solutions: AI Governance · Knowledge Graph