AI-Based Testing Interview — Q&A Kit¶
Simple, example-driven answers to the full AI-testing questionnaire (3 levels). Every question has a plain-English answer, one small Example, and a Remember hook.
| File | Level | Questions | Focus |
|---|---|---|---|
| 01-foundational.md | Foundational | 13 | LLMs, Generative AI, prompts, AI in testing, AI-assisted testing, risks |
| 02-intermediate.md | Intermediate | 17 | AI in frameworks, self-healing, hallucination, prompt testing/injection, context window, model drift |
| 03-advanced.md | Advanced | 17 | AI-first test strategy, agentic QA, autonomous agents, MCP + Playwright, KPIs, leadership/migration |
Total: 47 questions.
How to study¶
- Read an answer, then say it out loud in your own words.
- Memorize the Remember: hook for each — that's your recall anchor.
- For "have you used…" / "give an example" questions, swap in your real projects (LLM-eval work, Playwright/Selenium/RestAssured frameworks).
Related¶
Deeper versions of the LLM-eval, RAGAS, red-teaming and Vertex/Gemini topics live in your BCE AI-testing kit at ../bce-ai-testing/. This kit is the fast, plain-English pass over the exact questionnaire.
The one big idea (say this if unsure)¶
Traditional testing is deterministic — expected == actual. AI systems are probabilistic — the same input can give different output, and the model can be confidently wrong. So you shift from exact-match asserts to rubric scoring, semantic similarity, groundedness/hallucination checks, and pass-rate over many runs — and you keep a human in the loop to verify.