How we evaluate LLM applications in production

A practical evaluation framework for teams shipping copilots and RAG systems — without inventing a research lab.

MCMarcus Chen·May 28, 2026· 9 min read

Every team shipping an LLM product eventually asks the same question: how do we know it's actually working?

The answer is not a leaderboard score. It's a portfolio of evaluations, run continuously, that correspond to how your users experience the system.

We break evaluations into four layers. Layer one is unit-style: given a specific input, does the model produce the expected structured output? Layer two is retrieval quality: are we surfacing the right documents to ground the answer? Layer three is response quality, scored by a stronger model against a rubric. And layer four is human judgment on a stratified sample.

The most common mistake is over-indexing on the flashy last layer while skipping the boring first one. Structured-output evaluations catch 80% of regressions and cost you a few cents to run.

The second most common mistake is running evaluations only at release time. Production drift is real — user behavior, index freshness, and upstream model updates all shift the ground under you. Run your eval suite on a schedule, and gate deploys on it.

MC
Written by
Marcus Chen
Principal AI Engineer

Marcus builds production LLM systems. Formerly staff engineer at two AI-first startups and a hyperscaler.

More from Marcus

Get the next essay in your inbox.

One well-considered read from our practitioners, every other Friday.

Ready to amplify your data?

Book a 30-minute strategy call. We'll map your data landscape and identify the three highest-leverage moves.

See case studies