Trainer (Pod Lead) - Turing
I contributed to AI model training and assessment initiatives by designing and reviewing prompt–response scenarios that required structured tool usage (API calls, SQL queries, JSON schemas). Evaluated model outputs for correctness, logical consistency, policy compliance, and edge-case handling. Designed test cases that simulated real enterprise workflows; such as database state validation, multi-step tool invocation checks, and response verification. This is to ensure models produced executable, accurate outputs rather than plausible but incorrect answers.