Responsible AI Analyst
Accenture· India
Building evaluation frameworks, agentic systems, and production-grade AI tooling inside Accenture's Responsible AI practice. Focused on rigorous evaluation of LLM and RAG systems, developer experience for internal AI tooling, and translating research-grade evaluation techniques into reliable production pipelines.
Designed and shipped evaluation frameworks using Ragas, DeepEval, and custom Claude-based judges to benchmark RAG and agentic systems across faithfulness, relevance, and safety dimensions
Built agentic tooling on top of the Claude API — tool-calling loops, structured extraction, and reviewer-in-the-loop workflows — for internal teams evaluating client AI deployments
Owned end-to-end delivery of multiple client and internal engagements, from problem framing and dataset curation through dashboarding and stakeholder sign-off
Recognised with three merit hikes and a performance bonus within the first year for consistently high delivery quality
