Elevator Pitch
Build and deploy secure, production-ready MCP servers using Python and Gemini 3. Learn real-world patterns for authentication, scalability, and serverless deployment on cloud, taking AI backends from prototype to production.
Description
In this talk, we’ll explore how to build and deploy secure MCP servers using Python, powered by Gemini 3, and run them on serverless cloud infrastructure. Starting from a simple prototype, we’ll walk through the architectural decisions needed to make the system production-ready. Covering authentication, authorization, secrets management, observability, and cost-efficient scaling.
The session focuses on practical, real-world challenges faced when deploying AI services in production and demonstrates how Python developers can confidently ship secure AI backends without managing servers.
Why should you attend? Many AI projects fail when moving from demo to production due to security and operational complexity. This talk bridges that gap by showing how to build AI backends the right way - secure, scalable, and cloud-native—using tools Python developers already love.
Attendees will leave with a clear blueprint and best practices to design, secure, and deploy MCP servers on cloud using Python.
Notes
What this talk covers
Introduction to MCP (Model Context Protocol) and its role in agentic AI
Designing MCP-compliant servers using Python (FastAPI)
Integrating Gemini 3 securely into backend services
Authentication, IAM, and secrets management best practices
Deploying Python services on serverless cloud platforms
Scaling, logging, and observability for production workloads