Short entries on Development, Testing, AI, and Processes — real problems from real work, written up as I work through them.

I’m Yevhen (Eugene) Prodan, AI Quality Lead / AI Agent Architect for Quality Engineering at Luxoft, with 18+ years in test and project management for automotive software. This blog is where I think in public: I write up what I’m running into so I understand it better, and so others working on the same problems have something to compare notes against. More about me: About Me.

Browse by topic:

• Test Your Reviewer Like Code: A Red/Green Corpus for the Checklist Testing · AI
• Don't Tell the Model It Was Wrong — Update the Inspector's Checklist AI · Processes
• The Agent That Says "About Ninety Minutes" to Everything AI · Processes
• False Pass vs False Fail: The Verifier Error Nobody Prices Testing · AI
• The Afternoon I Lost to a Prompt: Technique Without Measurement Runs Blind AI · Development
• Quality per Token: The Tradeoff Nobody Puts on the Invoice AI · Processes
• A Batch Is Not an Increment: AI Moves the Bottleneck from Production to Verification Processes · AI
• Estimating the Triage Line When There's No Baseline to Size It From Processes · AI · Testing
• AI-Assisted Testing Adds a Validation Layer, It Doesn't Remove One Testing · AI