Agentic AI platform integrating Google Workspace, Notion, Todoist, and a PDF-based RAG system. Implemented computer-control capabilities through custom and official MCP servers, orchestrated by a LangGraph multi-agent architecture. Added sandboxed file access, tool execution, and human-in-the-loop approval workflows.
An enterprise-grade, autonomous AI orchestration system built with LangGraph, utilizing the Model Context Protocol (MCP) to seamlessly connect LLMs to local environments, vector databases, and cloud APIs. * Transitioned from a monolithic tool-calling agent to a distributed `MultiServerMCPClient` architecture.