InferenzInsights
Notes on building AI that works
Practical writing on production machine learning, LLM engineering and the ideas behind modern AI.
Production ML
Why most AI prototypes never reach production
The demo works and then nothing ships. Here are the real reasons AI prototypes stall, and a checklist for getting past them.
ML explainedSoftmax, explained: the idea behind the Inferenz logo
What softmax does, why every language model uses it, how temperature changes the output, and why it's the shape in our logo.
LLM engineeringRAG or fine-tuning? How to choose
A practical guide to deciding between retrieval-augmented generation and fine-tuning for your LLM application.