At AMBA System, I built an automated evaluation pipeline for a RAG-based Q&A chatbot using RAGAS and DeepEval. I designed a 150+ question Golden Dataset and regression, edge-case, and adversarial subsets to gate releases.
I also created a Promptfoo test suite for prompt injection and jailbreak scenarios, and automated safety regression runs in CI/CD. With LangSmith, I tracked evaluation runs, traces, and failures; I documented hallucination patterns and reported groundedness gaps for prompt and RAG tuning.
Earlier, I worked on the ThinkLink learning management system at AMBA System and payment and e-commerce projects at HCL Technologies and Computer Science Corporation (CSC). My testing work has included functional, regression, integration, compatibility, and mobile testing, and I hold ISTQB Foundation and Certified Scrum Master certifications.
