Skip to main content
maykaTech

Blog

Notes from the codebase, numbers included

What worked, what didn't, and what it cost — written by the practice leads who lived it. No thought leadership, no listicles, no AI-generated filler.

Laptop displaying code for a machine learning application
AI Engineering9 min read

The unglamorous checklist that gets LLM features into production

Most AI features die between the demo and the deployment. Here's the evaluation, guardrail, and cost work that separates the ones that ship — drawn from a dozen production launches.

Sofia Marchetti Head of AI Engineering

Read article