Litika Gopinathan

Architected and built a retail analytics platform powered by multi-agent LLM architecture enabling natural language querying over complex retail datasets. Implemented Agent-to-Agent (A2A) integration for seamless task delegation and inter-agent communication across heterogeneous AI agents. Designed and implemented multiple MCP servers for structured tool integration across AI agents. Built multi-agent orchestration pipelines using LangGraph with conditional routing, stateful workflows, and tool-use agents. Developed streaming response pipelines delivering real-time AI-generated insights. Built robust authentication and session management systems using OAuth2.0 and JWT for MCP Servers and Agents. Designed async FastAPI backends with dependency injection, background tasks, middleware, and structured error handling and logging.

Performed EDA and feature engineering on datasets; trained, validated, and evaluated ML models for retail forecasting and classification tasks. Built and deployed custom copilots via Microsoft Copilot Studio, integrating with the Power Platform and enterprise Microsoft ecosystem. Integrated AI agents into Microsoft Teams (Azure Bot Framework) and Google Chat for enterprise-wide deployment. Implemented PostgreSQL schemas with SQLAlchemy ORM and Alembic migrations for multi-tenant retail data.