~/blog/tag/ai
Notes from the field.
Writing on Django, Python, realtime systems and shipping business software that people use every day.
Evaluation and QA at Qoves
How AI output gets checked before a customer sees it: golden sets, tolerance gates, sampled review, and release rules.
Founding backend engineer at Qoves Studio
What owning the backend and AI platform of a health tech SaaS looks like in practice: queues, AI processing, reports, subscriptions.
From image analysis to a report customers trust
A computer vision pipeline is only half the product. The report the customer reads is the other half.
Evaluating AI pipelines before customers do
Evaluation sets, structured prompts, and monitoring: how we caught regressions without waiting for support tickets.
$ ./start_project.sh