Job Description
Design and execute test strategies for Gen AI models including LLMs image generators and multimodal systems
Create and maintain test cases and prompt libraries to evaluate model output quality eg relevance coherence creativity factual accuracy
Test model behaviour across different scenarios languages and user intents eg hallucinations bias toxicity safety risks
Perform manual and automated testing of Gen AI APIs chatbots and content generation tools
Validate input output behaviour prompt testing response evaluation token level analysis
Collaborate with engineering teams to implement prompt evaluation frameworks log monitoring and output scoring systems
Track model performance metrics BLEU ROUGE perplexity toxicity score etc and suggest areas for improvement
Conduct fairness bias and safety testing to ensure compliance with ethical and regulatory guidelines
Assist in evaluating AB experiments and fine tuning model behaviours
Document issues clearly and contribute to root cause analysis with dev teams.
