Gen AI Testing

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.