842 posts tagged with "ai-engineering"

Pull up the git blame on any year-old production system prompt and you will find something the engineering team is not ready to admit: the file has three authors, none of whom share a definition of what a "change" is. The engineer who refactored the instruction blocks last month logged the commit as "no functional change, just reordering for clarity." The product manager who reads the file once a quarter would describe the same diff as "you rewrote the voice — customers will notice." The ML engineer running the regression suite would call it "you broke few-shot example three, and the eval has been red ever since."

All three are right. The prompt is simultaneously code, spec, and hyperparameter, and every team that ships an AI feature long enough discovers that the file's commit history is a slow-motion three-way authorship dispute that CODEOWNERS does not capture and the diff viewer does not surface.

The deprecation email arrives on a Tuesday afternoon. The model your billing pipeline has depended on for fourteen months is now on a sixty-day timer. The prompt was tuned by an engineer who left in March. The eval suite hasn't been re-baselined since launch. The customer-success team is asking why "the AI feels different" on two enterprise accounts. Nobody put this on the roadmap, and nobody will own it cleanly, because in your org's mental model this is a one-off project — even though it is the fourth one this year.

Every team running an AI feature in production runs into the same realization within eighteen months: the foundation-model provider is operating on a deprecation cadence that the team did not plan for, and the team's migration response keeps being a reactive scramble triggered by a notification email. The fix is not a better playbook for the next migration — there are already plenty of those, and your team has probably written one. The fix is to stop treating migration as a project and start treating it as a recurring operational primitive. Put it on the calendar.

A customer asks your support agent how to do something. The agent retrieves three documentation chunks with high relevance scores, synthesizes a confident answer, and walks the customer through a five-step procedure that ends on a button that hasn't existed for four months. The customer files a ticket. The on-call engineer pulls the eval suite, finds it green, pulls the retrieval traces, finds them green too — the model didn't hallucinate, it faithfully quoted documentation describing a feature your product team renamed in the last quarterly release.

This is the failure mode I want to name: not a hallucination, not a retrieval miss, but a phantom inventory problem. Your retrieval corpus is a snapshot of a product surface that no longer exists. The vector store doesn't know the product changed. The eval suite doesn't know either. The only system that consistently catches it is the support ticket queue, and by the time a ticket is filed the customer has already been told to click a button that isn't there.

The team plans an eval suite the way they plan a service: failure modes inventoried, rubric drafted, sample size argued over, judge calibration scheduled. Then they file the rater capacity as a footnote — "we'll get the annotation team to grade a few hundred per week" — and ship the rest. Six weeks later the rater queue is at 4,300 items, eval velocity has collapsed to one judge-calibration cycle per month, and someone in a planning review says the quiet part out loud: nobody capacity-planned the humans.

Rater throughput is the binding constraint on eval velocity in any AI system that takes human grading seriously, and the discipline that treats annotation as an SRE problem rather than a recruiting one is the one that ships. A human reviewer processes 50–100 examples per hour at expert difficulty, and an expert annotator caps out around 500–1,000 examples per week — those numbers are not a recruiting problem to be brute-forced with headcount. They are an operational property of the eval system that has to be modeled and budgeted the way you model database IOPS.

A user opens your chat, asks a question, and gets back a response your eval suite would score 4.6 out of 5. Then they ask the same question with different words. Same answer. Same score. They try once more, this time with the kind of hedging language people use when they suspect the machine isn't listening — "what I'm actually trying to do is…" — and then they close the tab. From the model's perspective, three clean Q&A turns. From the dashboard's perspective, an engaged session. From the user's perspective, a product that failed them three times in a row and won't be opened again.

This is the failure mode per-turn evaluation cannot see. Each individual turn looked correct in isolation. The judge gave a thumbs up. The hallucination detector stayed quiet. The relevance score was high. And yet the conversation, as a whole, did not resolve anything — and that's the unit the user was actually evaluating you on.

A team I talked to recently sold their legal-AI product into a Fortune 500 in-house counsel office and added one line to their system prompt: "every factual claim must include an inline citation to the retrieved source." The product roadmap allocated a 5% buffer on their token budget for the new behavior. Sixty days after the regulated tenant went live, finance flagged a 34% jump in monthly inference spend. Nobody had broken the product. Nobody had shipped new features. The compliance requirement that closed the deal also quietly rewrote the unit economics underneath it.

This is the retrieval citation tax, and almost every RAG system serving a regulated industry — legal, healthcare, finance, audit-bound enterprise — eventually pays it. The tax is structural, not a bug. It comes from the way citation discipline forces the model into a different generation regime, and it shows up nowhere on the procurement spec the customer signed.

The fastest way to discover that your AI team has no eval discipline is to ask three engineers, in separate Slack DMs, "did your last prompt change improve quality?" — and watch them answer yes, all three of them, with three different numbers, against three different slices, on three different laptops, none of which is reproducible by anyone else in the room. That isn't an evals problem in the textbook sense. The textbook says you don't have evals. The reality is worse: you have too many evals, each of them privately owned, each of them measuring something real, and none of them rolling up into a single number the org can plan against.

This is the shadow eval anti-pattern, and most AI teams ship with it for longer than they admit. It looks productive — every engineer has a notebook, every PR comes with a screenshot of a pass rate, every standup mentions a "win on the long-tail slice" — and it survives quarterly reviews because the bar for "we do evals" is so low that running anything counts. But the org has no signal. Leadership cannot tell whether last month's three prompt edits moved the product forward or sideways, because the three engineers measured against three private slices and stopped tracking the previous baseline the moment they switched files.

Open the system prompt of any AI feature that has been in production for more than nine months. Scroll past the role description, past the formatting rules, past the safety guardrails. Stop at the block titled <examples> or ## Examples or whatever your team called it the day someone copied the first three good Slack threads into a code block. Read them. There is a 60% chance at least one of them references a feature that has been renamed, a button that no longer exists, or a workflow the product manager quietly killed two quarters ago.

The decay is not visible from the eval dashboard. The eval scores are green. They have been green for months. They are green because the eval set was authored against the same product surface the few-shots reference, and the two have aged together in lockstep. The model is performing a flawless impression of last year's product, on a test set that grades it for being faithful to last year's product, while real users interact with this year's product and quietly tolerate the resulting confabulations. This is the half-life nobody puts in the LLMOps roadmap.

Every team running an AI feature in production is sitting on the highest-signal eval dataset they will ever have, and they are not using it. The dataset is in Zendesk. Or Intercom. Or Freshdesk, or Help Scout, or whatever queue the support team lives inside. The tickets that get filed there describe the exact failure modes the model produced in front of a paying customer — wrong tone, wrong tool call, wrong policy, hallucinated capability, leaked context. Each one is a labeled negative example, hand-written by the user who experienced the failure, often with reproduction steps and a sentiment annotation attached for free.

The eval suite, meanwhile, lives in Git. It was hand-written by whichever engineer set it up six months ago, and it has accumulated maybe fifty cases since. The intersection between "things the eval suite covers" and "things that actually break in production" is a Venn diagram with a thin sliver of overlap and two large, mutually ignorant lobes.

Your eval suite ran green at 2 AM PT on a quiet provider. QA smoke-tested at 11 PM the night before launch. The feature goes live, and by Tuesday at 10 AM Eastern your p95 is 40% higher than the dashboard you signed off on, your agent is dropping the last tool call in a six-step plan, and your support inbox is filling with tickets that all sound the same: "the AI was weird this morning." Nobody is wrong. The model is also not wrong. The eval set is wrong — it never saw a saturated provider, so it has no opinion on what the feature does when the queue depth triples and the deadline budget collapses.

Provider load is not a latency problem with a quality side effect. It is a distribution shift in the inputs your model and your agent loop receive, and you have built every quality signal you trust on the wrong half of that distribution. The fix is not a faster region or a better model. The fix is to stop pretending your eval harness is sampling from the same world your users are.

A new optional parameter goes into a tool description on a Tuesday. The change is small — six lines in the diff, no breaking signature change, no callers updated, no eval cases touched. The PR description says "adds support for an optional language filter to the existing search tool." Two reviewers approve. It ships.

A week later, the cost dashboard shows that the search tool is being called eighteen percent more often than the prior baseline. Latency on the affected agent has crept up by roughly the same proportion. Nobody can point to a single failing eval. The new parameter, when used, behaves correctly. The new parameter, when not used, doesn't matter. And yet the planner has clearly changed its mind about when to reach for this tool — and the eval suite, which grades tool correctness, has nothing to say about a shift in tool frequency.

Open the system prompt of any AI feature that shipped in the last six months and read it side by side with the PRD that authorized it. You will find two documents arguing with each other. The PRD says "the assistant should be helpful but professional, avoid making things up, and gracefully decline if it can't answer." The system prompt says "You are an AI assistant. Be concise. If you are unsure, say 'I don't know.' Never invent facts." The PRD takes a page. The prompt takes nine lines. The gap between them is where every behavioral bug you shipped this quarter lives.

The convenient fiction is that the prompt is an "implementation detail" of the PRD. The actual relationship is the opposite. The prompt is the contract the model executes; the PRD is a draft of that contract written in a language the model does not speak, by an author who never compiled it. Every PRD for an AI feature is an untested prompt. The team that admits this and runs the PRD through an eval before sign-off ships a feature with one fewer source of post-launch surprise.