Ankur and Dulan Agentic AI Lab is a collaborative engineering lab dedicated to bridging the gap between unstructured real-world data and actionable agentic intelligence.

We believe that for AI agents to be truly effective, they require high-fidelity "senses" and tools. Our work focuses on:

• Async MCP (Model Context Protocol) Server: Developing robust MCP (Model Context Protocol) server to give your AI agents, structured access to specialized tools and data.

• Structure: Utilizing advanced Document Intelligence to extract and map complex hierarchies from unstructured data into machine-readable knowledge bases.

• Perception: Implementing real-time Computer Vision systems that transform raw camera feeds into state-aware agentic inputs.

By combining Production-Grade MLOps (Ankur) with Advanced Computer Vision (Dulan), we build systems that don't just "think"—they perceive and act.