Practical AI, from someone who ships it.
Advisory, hands-on delivery, and training for teams that want machine learning and generative AI to actually work in production, not in slideware.
Kaggle Grandmaster · 3 books · PhD · 20+ years in quant finance & ML
How I can help
From a second opinion on your AI strategy to building the system alongside your team. Pick the level of involvement that fits. Always vendor-neutral: I don't resell anything, so the advice is the product.
GenAI & ML Advisory
Know if it's real before you spend. An independent assessment of your AI plans: strategy, architecture, evaluation. You get a straight answer to "is this real or hype for our use case?" and a roadmap your team can execute.
Hands-on Delivery
Ship the system, not the demo. I build alongside your team: RAG, agents, LLM evaluation, time-series forecasting. It ends with working software in production, in your stack, owned by your people.
Training & Workshops
Make your team self-sufficient. Workshops and self-paced courses (GenAI Code, AI Agents) that turn engineers into people who ship AI. See what participants say.
Speaking & Tutorials
Technical content, zero fluff. Conference talks, brownbags, and the "Time Series with Konrad" tutorial series. Material that respects your audience's intelligence.
Written the book on it. Three, actually
Practical machine learning, straight from the competition arena.
The Kaggle Book
The bestselling guide to competitive machine learning: data analysis, modeling, and the mindset that wins.
View on Amazon →
The Kaggle Workbook
Hands-on exercises and detailed solutions that turn the theory into muscle memory.
View on Amazon →
Machine Learning Using TensorFlow Cookbook
Practical recipes for building real ML systems with TensorFlow 2.
View on Amazon →
From quant finance to GenAI
I've spent 20+ years turning data into decisions: a PhD in credit risk modeling at the Vrije Universiteit Amsterdam, quant roles at ING, RBS, and ABN AMRO, leading data science at eBay, building generative AI systems at IKEA, and now heading AI at WAIR. The problems along the way ranged from high-frequency trading and credit risk to commodity prices and predictive maintenance. Always the full lifecycle, from business question to production model.
Along the way I became a Kaggle Grandmaster, co-authored three books on practical machine learning, and ran TNG Quant Consultancy since 2010, delivering ML solutions and training on a project basis.
The common thread: I care about what works, not what demos well. That's the perspective I bring to every engagement.
Selected publications
- "Modeling portfolio defaults using Hidden Markov Models with covariates" · Econometrics Journal, 11 (2008)
- "Quantile Forecasting for Credit Risk Management using Possibly Mis-specified Hidden Markov Models" · Journal of Forecasting, 27 (2008)
- "Tail dependence of grouped skewed t-distributions" · Statistics & Probability Letters, 78 (2008)
- "On copula based modeling of economic time series" · Free University Research Report WI 17 (2004)
- since 2010Owner, TNG Quant Consultancy
ML solutions & training on a project basis - 2026 – presentHead of AI
WAIR · AI agents for fashion retail: forecasting, agentic workflows - 2023 – 2026Principal Data Scientist
IKEA · GenAI systems, evals, multimodal agents - 2017 – 2023Lead & Staff Data Scientist
eBay · search ranking, recommender systems, LLM moderation - 2016 – 2017Lead Data Scientist
Semiotic Labs · predictive maintenance - 2010 – 2016Quant roles
ING, RBS, Solvency II modeling for Aegon, Delta Lloyd & others - 2009PhD, credit risk modeling
Vrije Universiteit Amsterdam · MSc in Mathematics, University of Warsaw
Let's talk about your AI problem
Working with me is deliberately simple:
Free 30-minute call
You describe the challenge. No prep needed, no pitch decks.
Short written assessment
What I'd do, how long it takes, what it costs. Honest, including "you don't need me for this."
Fixed-scope engagement
Advisory, delivery, or training, with clear deliverables and an end date.