Engineering notes
Practical writing on what actually works when you take AI/ML systems to production. No hype, no buzzword soup — just the patterns we use on real client engagements.
-
How Much Does a Custom LLM Fine-Tuning Project Cost? A Realistic 2026 Breakdown
An honest cost breakdown — data, compute, engineering, ongoing ops — plus the three project-size bands enterprise buyers actually see, and when fine-tuning is the wrong answer.
-
RAG in Production: 7 Failure Modes We Fixed for Enterprise Clients (and How)
The seven mistakes that turn a beautiful RAG demo into a production-grade liability — and the architectural patterns that fix each one.