Managed annotation and evaluation teams for high-quality CUA datasets, established SOPs, rubric documentation, and calibration protocols. Delivered daily/weekly updates, workload balancing, and risk management across concurrent projects.

Led NVIDIA-sponsored Computer Use Agent (CUA) data programs, managing 300+ distributed team members, ensuring data quality, process governance, and stakeholder communication for next-gen AI agent training.

Designed multi-agent layer (LangChain/LangGraph), authored AI explainability reports, conducted prompt engineering and LLM tuning, and tailored AI insights for multiple personas in regulated environments.

Developed Adaptive Process Orchestration Agentic AI web app for BFSI, integrating multi-agent frameworks and explainability reporting, leveraging Claude Code for accelerated development.

Championed Microsoft 365 Copilot adoption, co-designed upskilling roadmap, optimized Databricks+PySpark pipelines, and delivered production models with measurable business impact.

Owned end-to-end AI/ML stack, led enablement for Data & Analytics team, architected LLM+RAG platform on GCP, delivered scalable HR analytics models, and improved team throughput by 30%.