Anna Kriuchkova

Case files

Real problems, briefed properly

Each file is something that actually went wrong, or needed building, in the AI system that runs our house. They are told the way I would brief them at work: the problem, the evidence, the fix, and what I do differently now.

A note on credit. The system is a joint project with my husband, and AI coding agents write most of its code. What is mine in these files is the part that matters to an employer: the question asked, the evidence demanded, and the decision made.

Case file 01

The rock that was a bird

One camera logged 179 birds in three days. The checking model agreed with every one of them. Both were wrong, for different reasons.

  • computer vision
  • false positives
  • A/B testing

Case file 02

The camera detects. A second model decides.

Rebuilding the notification feed so that a wrong label never reaches a phone, and finding a silent failure that had been hiding a total outage.

  • computer vision
  • silent failure
  • verification

Case file 03

Twenty times faster by moving one thing

Camera vision took up to two minutes a frame. The fix was not a bigger model or more memory. It was asking the right question about the hardware.

  • performance
  • model comparison
  • migration

Case file 04

The dryer is done, says the television

Small mercies, engineered properly: laundry announcements, a doorbell that takes over the screen, and an intercom that answers back.

  • requirements
  • home automation
  • knowing when to stop

Case file 05

The search index that was quietly rotting

My study desk's search depended on a paid service with a shutdown date. Nothing looked broken, and nothing would have looked broken afterwards either.

  • PostgreSQL
  • vector search
  • verify by count

Case file 06

The rules the AI cannot argue with

Read-only by default, every change briefed in plain language first, and a guard that blocks destructive commands no matter how politely they are phrased.

  • AI governance
  • change control
  • audit

"A wise man, therefore, proportions his belief to the evidence."

Contact

Have data that needs a second look?