Backend developer. I explore how AI changes the SDLC, the roles of engineers, and eventually - the toolset.

What I build: usual backend applications, day by day. What I am really into: agentic workflows, deterministic pipelines and dark factories, where agents do the routine work and humans stay for final decisions.

What I experiment with: how SDLC changes when AI agents write, test, review, fix, explain the code (and whatever you use AI nowadays). Which parts stay human, which parts can be fully automated, and what breaks on the way.

Grounding: is my favorite. An agent without “ground” is only guessing and hallucinating. And tests are very interesting part of the grounding, believe it or not. I work on all layers, unit, integration, API, e2e and manual, to shift feedback left as much as possible; and also push tests down, into the fastest layer. So both developers and agents get feedback literally in minutes (my goal).

From the backend dev point of view, topics I care about and want to discuss:

• Bug reports. Without video. Evidence instead: reproduction ID, traces, logs, requests, etc.
• Exploratory session into a concise bug report with everything the dev team needs, and with minimal amount of time/effort
• Code coverage for manual/API/e2e testing
• Test run evidences for highly regulated industries
• Manual test as a regression test
• Test pyramid rebalanced by risk and coverage
• Tests as grounding for AI agents

Happy to discuss, argue and learn from testers who see this from the other side.