Speak at Bengaluru Systems meet-up

Elevator Pitch

Discover how to turn the Gemini CLI into a powerful Data Engineering tool. This session provides a hands-on look at using the Model Context Protocol (MCP) to connect Gemini CLI directly to Apache Pinot, enabling natural language querying and agentic workflows over streamable HTTP.

Description

With the rise of “vibe coding” and agentic workflows, the CLI has returned as the interface of choice for data professionals. The Gemini CLI has emerged as a robust client for the Model Context Protocol (MCP), supporting advanced features like streamable HTTP, stdio transport, and OAuth.

In this talk, we will bridge the gap between Generative AI and Big Data systems. We will demonstrate how to leverage Gemini CLI to interact with Apache Pinot, a real-time distributed OLAP datastore.

By moving beyond simple chat interfaces and integrating directly with data infrastructure, we empower Data Engineers and Scientists to query and analyze massive datasets using natural language.

Key Takeaways

1. Architecture Analysis: Understanding the architecture of Gemini CLI as an MCP Client.

2. Protocol Deep Dive: A comparison of MCP transport types: Streamable HTTP vs. stdio.

3. Practical Guide: Step-by-step configuration of Gemini CLI to talk to a local Apache Pinot cluster.

4. Live Demo: Querying Pinot data using natural language prompts directly from the terminal.