a
adamstanai

Adam Stan

@adamstanai

AI QA Engineer LLM, RAG Voice Agent Testing

Poland
English, Polish
About me
I help companies find critical quality issues in AI products before their users do. I'm a QA Engineer specializing in LLM applications, RAG systems, AI chatbots and voice agents. I test response quality, hallucinations, groundedness, context retention, conversation flows, STT/TTS, edge cases, APIs, end-to-end workflows and guardrails. I deliver structured test scenarios, PASS/FAIL evaluations, severity ratings, reproducible bug reports, evidence and actionable recommendations to make AI systems reliable for real users.... Read more

Skills

a
adamstanai
Adam Stan
Offline • 

See my services

Software Testing
I will professionally test your ai chatbot, llm or rag application
From$40 / projectMore details

Work experience

AI QA Engineer

Zabka Polska sp • Full-time

Apr 2026 - Present • 5 mos

Quality assurance for conversational AI and voicebot systems. • Testing LLM-based conversational workflows • Evaluating response quality, correctness and edge cases • Voicebot testing including STT, TTS, interruptions and conversation flows • Testing multi-turn conversations and context handling • Regression and exploratory testing • API and end-to-end validation • Designing structured AI evaluation scenarios • Investigating production-like failures and unexpected AI behavior • Collaborating with developers and AI teams on quality improvements

QA Engineer / Test Automation Engineer

Transitions • Freelance

Mar 2022 - Mar 2026 • 4 yrs

Software QA Engineer with experience testing complex enterprise and energy trading systems. • Manual and automated software testing • End-to-end and regression testing • API testing using REST and SOAP • Test automation with Playwright and Cypress • SQL and database validation • Integration testing • Kafka-based system testing • Bug analysis and reporting • CI/CD environments including GitHub Actions and Jenkins Built and maintained automated regression suites that significantly reduced manual regression testing time.