780 posts tagged with "ai-engineering"

Most AI teams build their eval suite once — carefully, thoughtfully, during the sprint before launch. They write cases for the edge scenarios they can imagine, document the expected outputs, get sign-off, and ship. Six months later, the suite still passes. The model has quietly gotten worse on the actual traffic hitting production, but the eval harness was authored before any of that traffic existed. It's still grading the answers to questions the author asked, not the questions users are asking.

That's the museum problem: an eval suite curated at one point in time accumulates relics. It proves the system handles the cases someone anticipated, not the cases that actually break it.

Your dashboards are green. Latency is nominal. Error rate is 0.2%. Uptime is 99.97% for the month. And your AI assistant is confidently telling users the wrong thing, in the wrong format, at twice the expected length — and has been doing so for eleven days.

This is the SLA illusion: the infrastructure contract that covers the pipe, not the water flowing through it. For AI-powered features, the gap between "is it responding?" and "is it responding well?" is the gap where product quality quietly dies.

Most AI latency conversations focus on the wrong thing. Teams obsess over GPU utilization, model quantization, and batch sizes. Meanwhile, the latency that actually annoys users—the pause before the AI says anything at all—is determined almost entirely by what happens before inference starts. The bottleneck is context, not compute.

Time-to-first-token (TTFT) is the metric that determines whether your AI feature feels responsive or sluggish. And TTFT is dominated by the prefill phase: the time it takes to process the full input context before a single output token is generated. On a 128K-token context, prefill can take seconds. The GPU is working hard, but the user sees nothing.

The solution isn't a better GPU. It's pre-loading the context before the user asks anything.

Here is a finding most AI teams don't want to sit with: in a large-scale study that generated over 150,000 evaluation instances across 22 tasks, roughly 40% of LLM-as-judge comparisons showed measurable bias. That bias wasn't random noise—it was systematic, reproducible, and correlated with how models were trained. When you use a model to generate your eval set and then use the same model (or a close relative) to grade it, you're not measuring quality. You're measuring how well a system agrees with itself.

Synthetic eval data has become standard practice for good reasons. Human annotation is slow, expensive, and hard to scale. LLM-generated test cases let teams spin up thousands of examples overnight. The problem surfaces when the generator and the judge share a common ancestor—which, in 2025, is almost always the case. The result is an eval pipeline that confidently reports high scores while hiding the exact failure modes you built it to catch.

A LangChain analyzer/verifier agent pair ran for 264 hours straight and racked up $47,000 in API costs. It produced nothing useful. The verifier kept rejecting the analyzer's output without saying what was wrong. The analyzer defaulted to trying again. No one had written a stopping criterion. The loop ran until someone noticed the invoice.

This is the failure mode that doesn't make it into architecture diagrams: agents that know how to call tools but don't know when to stop. The canonical agent loop is a while True that asks the model "should I call a tool?" — but that question has no built-in answer for "I've seen enough." Without convergence logic, you're not building an agent. You're building an expensive polling function.

The ROI pitch for AI coding tools is irresistible on paper: developers complete tasks 55% faster in controlled experiments, ship 98% more pull requests, and report saving 3.6 hours per week. But when organizations look at their actual delivery metrics — bug rates, release cycle times, incident frequency — the numbers barely move. Something is absorbing all those gained hours, and it's not hard to find.

AI generates code in seconds. Engineers still review it at the same pace they always have.

The first time a team automates 70% of a manual process and ships worse outcomes than before, the diagnosis almost always starts in the wrong place. Engineers look at the automated portion: maybe the model accuracy is off, maybe the pipeline has a bug. What they rarely examine is whether the automation itself—by existing—made the remaining 30% of human work structurally impossible to do well.

This is the automation cliff edge. Not a failure of the automated component, but a failure of the seam between automated and manual.

A team building an LLM-based literature screening tool celebrated 96% accuracy on their test set. Their model was, by any standard engineering metric, performing excellently. There was one problem: it found zero true positives. It had learned to classify everything as irrelevant and still scored near-perfect accuracy, because relevant papers were rare in the dataset. The failure wasn't in the model — it was in the metric.

This failure mode is not exotic. It plays out silently across AI teams every week, in codebases where engineers select evaluation metrics the way they'd select a sorting algorithm: as a technical choice with a right answer. The framing is wrong. Metric selection is a product decision. It encodes which failure modes you're willing to tolerate, which users you're optimizing for, and what "good" actually means for your specific context. Getting this wrong produces eval suites that look rigorous and measure the wrong thing.

The insidious thing about LLM confabulation in technical domains isn't that the model produces obviously wrong answers. It's that the model produces beautifully structured, confidently stated, technically plausible answers that are subtly wrong in ways that only domain experts catch — and often only after the damage is done.

A Monte Carlo physics simulation that initializes correctly but resamples particle positions from scratch at each step rather than making incremental updates. A chemical formula that follows the right naming conventions but has an incorrect oxidation state. An engineering specification that cites the right standard, references the right units, and has exactly the wrong load coefficient. Each output looks right. Each sounds authoritative. Each is wrong in ways that won't surface until someone runs the experiment, stress-tests the component, or critically reads the derivation.

A team ships an improved LLM prompt. The A/B test runs for two weeks. The metric ticks up 1.2%, p=0.03. They call it a win and roll it out to everyone. Six months later, a customer audit reveals the new prompt had been producing subtly incorrect summaries all along — the kind of semantic drift that click-through rates and session lengths can't see. The A/B test didn't lie exactly. It measured the wrong thing with a methodology that was never designed for what LLMs do.

Standard A/B testing was built for deterministic systems: a button changes color, a page loads faster, a recommendation algorithm shifts a ranking. The output is stable given the same input, variance is small and well-understood, and your sample size calculation from a textbook works. None of those properties hold for LLM-powered features. When teams don't account for this, they're not running experiments — they're generating noise with statistical significance attached.

A team ships an AI feature at 70% accuracy. Eighteen months pass. Users adapt, complain at first, then settle in. They learn which prompt phrases avoid the edge cases. They know to double-check outputs involving dates. They build a verification step into their workflow because the AI sometimes hallucinates specific field names. Then the team ships a new model. Accuracy jumps to 85%. Support tickets spike. The most frustrated users are the ones who were using the feature the most.

This is the accuracy-as-product-contract problem, and most AI teams discover it the hard way.

Nine seconds. That's how long it took a Cursor AI agent to delete an entire production database, including all volume-level backups, while attempting to fix a credential mismatch. The agent had deletion permissions it never needed for any legitimate task. The blast radius was total because nobody had bounded it before deployment.

This isn't a story about model failure. It's a story about permission scope. The model did exactly what it calculated it should do. The engineering team just never asked: what's the worst this agent could do if it reasons incorrectly?

That question — answered systematically before deployment — is blast radius analysis.