https://hendorf.com/en/blog/ 2026-04-26T21:16:07+02:00 Alexander C.S. Hendorf https://hendorf.com/en/blog/stop-waiting-start-shipping/ 2026-04-26T21:16:07+02:00 2026-04-27T00:00:00Z

Sebastian Raschka in a fireside chat: why 99.9% of companies do not need to train their own base model — and where Europe's real sovereignty is built.

Alexander C.S. Hendorf https://hendorf.com/en/blog/agentic-ai/ 2026-04-26T23:35:34+02:00 2025-09-15T00:00:00Z

Three months of coding agents in PyCon DE & PyData operations: where they hold up, where they fail unexpectedly, and why governance matters more than tool choice.

Alexander C.S. Hendorf https://hendorf.com/en/blog/cross-pollination-ai/ 2026-04-26T23:35:34+02:00 2025-08-15T00:00:00Z

AI is permeating every layer of the organisation. The companies that win in the era of autonomy aren't the ones with the best single technology — they're the ones with the best cross-pollination across boundaries.

Alexander C.S. Hendorf https://hendorf.com/en/blog/enterprise-ai-open-source/ 2026-04-26T23:35:34+02:00 2025-04-20T00:00:00Z

From a fireside chat at PyCon DE & PyData 2025 with Merck, Quoniam and Explosion/spaCy: open source as the pragmatic enterprise default, standardisation as the unsexy high-leverage investment, and why the Mittelstand is faster than the usual narrative claims.

Alexander C.S. Hendorf https://hendorf.com/en/blog/data-quality-assurance/ 2026-04-26T21:16:07+02:00 2025-03-15T00:00:00Z

Data quality decides whether AI systems hold up in production. Patterns from enterprise mandates, briefed for science-spinout founders so they avoid the same expensive mistakes.

Alexander C.S. Hendorf https://hendorf.com/en/blog/overfitted-promises/ 2026-04-26T23:35:34+02:00 2025-03-15T00:00:00Z

Today's AI coding tools work more like alchemy than like established science — and that's not a verdict, it's a stage. Where they pay off, where they fail, and the heuristics that hold up in the meantime.

Alexander C.S. Hendorf https://hendorf.com/en/blog/warum-ki-projekte-scheitern/ 2026-04-26T23:35:34+02:00 2024-11-15T00:00:00Z

AI projects rarely fail on the technology. They fail on people, organisations and missing governance — three acts of failure, and the levers that successful programmes pull instead.

Alexander C.S. Hendorf https://hendorf.com/en/blog/economics-of-prediction-machines/ 2026-04-26T23:35:34+02:00 2020-11-15T00:00:00Z

AI is not IT — it is R&D. The companies that win as prediction costs collapse are the ones that treat prediction capacity as strategy, not as a technology purchase.

Alexander C.S. Hendorf https://hendorf.com/en/blog/ki-fuer-entscheider/ 2026-04-26T23:35:34+02:00 2020-03-15T00:00:00Z

AI is not a product you buy — it is an organisational and technological capability you build. Six years of LLM acceleration haven't changed that; the human work is still where most programmes break.

Alexander C.S. Hendorf https://hendorf.com/en/blog/deep-learning/ 2026-04-26T23:35:34+02:00 2020-02-11T00:00:00Z

Two years of hands-on PyTorch experiments — style transfer, text generation, speech synthesis — and the unglamorous lessons about data quality, the 90% trap and the real gap between lab and production.

Alexander C.S. Hendorf https://hendorf.com/en/blog/agile-analytics-enterprise/ 2026-04-26T23:35:34+02:00 2018-03-15T00:00:00Z

A 2018 insurance case study: out of proprietary silos, into a Python-based open-source stack — with the change-management discipline that the technology alone cannot provide.

Alexander C.S. Hendorf