Skip to Main Content
Talk Intermediate

Build Python AI Agents that can talk to any system in minutes using secure MCP

Proposal status is Rejected
Session Description

What if your Python AI agent could connect to Jira, SQL, or SAP in minutes—securely and at scale? In this talk, you'll Plug Into the MCP Matrix and discover how open-source MCP servers and registries can supercharge your LLMs with real-world integration superpowers. Learn how to create your own secure MCP registries and use mcp package managers like fluidmcp using a newly open opensource mcp package manager https://github.com/Fluid-AI/fluidmcp to instantly spin up FastAPI endpoints that bridge AI with complex systems—no boilerplate, no headaches. With live demos and battle-tested insights this session is your blueprint for turning AI agents into real-world action engines

Key Takeaways

Plug Into the MCP Matrix: Securely Hack the Stack and Integrate AI with Anything in Minutes

Session Outline:

  1. Introduction (5 min)

The big problem: LLMs are smart—but disconnected from real systems

What is MCP (Model Context Protocol) and why it matters

  1. The Vision of MCP Servers & Registries (8 min)

Why traditional integrations are slow, insecure, and developer-unfriendly

How MCP abstracts real-world systems (Jira, SQL, SAP, CRMs) into secure, modular endpoints

Introducing open-source MCP Registries and the fluidmcp package manager

  1. Under the Hood: Building with Python + FastAPI (8 min) using the newly open sourced fluidmcp package manager

Anatomy of an MCP server: schema, auth, secure token flow

Live walk-through: spinning up an MCP integration with FastAPI

How registries help you version, publish, and control access

  1. Live Demo: AI Agents Taking Real-World Actions (10 min)

Connecting an AI agent (LLM) to Jira via MCP

Use case: Create a ticket, assign, and comment—autonomously

Connecting to a SQL DB securely using MCP

Showcasing secure logging, audit, and registry-level access controls

  1. Scaling Securely with Open Registries (6 min)

Hosting MCP registries: self-hosted vs enterprise

Security best practices: token auth, action controls, and sandboxing

How to plug your own backend tools into MCP

  1. The Dev Superpower (3 min)

What took weeks now takes minutes—with security baked in

How Python devs can now build true agentic systems

Invitation to contribute to the open MCP registry ecosystem

References

http://fluidmcp.com
https://github.com/Fluid-AI/fluidmcp

Session Categories

Introducing a FOSS project or a new version of a popular project

Which track are you applying for?

Compilers, Programming Languages and Systems Devroom

Speakers

Abhinav Aggarwal
Abhinav Aggarwal Founder and CTO | Fluid AI

Abhinav Aggarwal is one of the founders of Fluid AI (www.fluid.ai). Abhinav dropped out of the MBA at India’s number 1 insititute (Indian School of Business) to found his company, Fluid AI. He has been working in deeply in AI research and product for organizations for the last 15 years. He is a part of the Forbes 30 under 30 Asia list and the Fortune 40 under 40 India list. Abhinav is a Chartered Accountant. However, he has always had a passion for coding and is a self-taught coder. He began his journey when he emerged as one of the winners at the world’s biggest 24-hour coding competition, the TechCrunch Disrupt Hackathon in New York in 2012. His case study about him and Fluid AI is now being taught at Harvard Business School.

He was recently featured on the cover of Forbes Magazine. He has also worked with and met Warren Buffett to launch the world’s first A.I. magazine cover for the Forbes 100th year

anniversary. He was also the person behind the world’s first book written by an AI to become an international bestseller. He is a guest faculty at ISB Hyderabad. In addition, he has launched the world’s first People-less banking branches powered by AI for Bank of America.

Revolutionizing the Artificial Intelligence space, his company Fluid AI provides customer focused experiences as well as decisioning products. Their clients include Bank of America, Barclays, Societe Generale, Royal Bank of Canada, Capgemini, Vodafone, Toyota, Intel, Rolls Royce, Government of Mauritius, Electricite de France, Daimler Mercedes etc.

Fluid AI has been listed by Forrester in its latest global AI report as the ‘Top Vendor for Deep Learning’ and by Gartner as a ‘Cool Vendor for AI’. IBM has selected them as the best ‘Deep Technology Startup’. Frost and Sullivan has awarded them as ‘Artificial Intelligence Product of the Year’. Fluid AI has also been a part of the Mastercard Start Path program.

https://www.linkedin.com/in/abhinav-fluidai
Raghav Aggarwal
Raghav Aggarwal Founder and MD | Fluid AI

Raghav Aggarwal is the founder of Fluid AI (www.fluid.ai). He dropped out of IIM Ahmedabad (Ranked No.1 in India) to start the company. He was recently featured on the cover of Forbes Magazine and is a part of the Forbes 30 under 30 Asia list and the Fortune 40 under 40 India list. Raghav is a Chartered Financial Analyst. However, he has always had a passion for coding and is a self-taught coder. He was one of the winners at the world’s biggest 24-hour coding competition, the TechCrunch Disrupt Hackathon, in New York in 2012. He was also behind the world’s first book written by an AI to become an international bestseller. He is a guest faculty at ISB Hyderabad. In addition, he helped launch the world’s first People-less banking branches powered by AI for Bank of America.

He has also worked with Warren Buffett to launch the world’s first A.I. magazine cover for the Forbes 100th year anniversary. Fluid AI was also telecasted by the BBC World News across 70 countries showcasing the work being done in the Artificial Intelligence space.

The vision of Fluid AI is to be the first company to pass the Turing Test—building an AI that can converse with humans without them realizing there is a machine at the other end. The company has 3 patents granted in the USA.

https://www.linkedin.com/in/trutech