How Tinybird Built a Full Data Dev Agent with Gemini

Google Cloud · Intermediate ·💻 AI-Assisted Coding ·3mo ago

Key Takeaways

Tinybird built a full data development agent with Gemini, evolving from a simple SQL generator to a co-pilot for data development, powered by Gemini's advanced natural language processing and large context window handling.

Full Transcript

I'm Javi, co-founder Tinybird. At Tinybird, we want to enable our users [music] to build and ship their data products and user-facing analytics much faster. Two key features in our platform that allow for advanced data exploration are powered by Gemini. Explorations allows our [music] users to have a natural language conversation with their data. They can ask questions, [music] request insights, and create visualizations, all without writing a single line of code. Gemini translates their questions [music] into complex queries, runs them, and presents the results in an understandable [music] way. Gemini is also behind our NCP server functionality. Our NCP [music] server enables our users to build their own custom data agents. This empowers [music] them to create highly specialized AI-driven applications on top of Tinybird. So, they can, for example, offer customized insights [music] and automations to their end users. Gemini's advanced ability to handle very [music] large context windows is a game-changer. It makes a difference for developers that need to work with large code bases or do deep data analysis. It helps them to go faster and avoid workarounds.

Original Description

Tinybird is a real-time data platform built on ClickHouse, and Tinybird Code is its AI-powered development agent. In this video, Tinybird cofounder Javi Santana shares how the team used Gemini to evolve Tinybird Code from a simple SQL generator into a full data development co-pilot. Today, Tinybird Code collaborates with developers across the entire data lifecycle—producing cleaner SQL, reducing errors, and cutting debugging time. Gemini also powers Tinybird’s exploration features, enabling natural language conversations with data to generate insights, queries, and visualizations without writing code. Beyond development, Gemini drives Tinybird’s MCP Server, which allows customers to build custom data agents and AI-powered applications directly on top of Tinybird.
☆ Save to playlist Watch on YouTube ↗
Sign in to unlock AI tutor explanation · ⚡30

Tinybird used Gemini to build a full data development agent, leveraging its advanced natural language processing and large context window handling to evolve from a simple SQL generator to a co-pilot for data development.

Key Takeaways
  1. Identify the need for an AI-powered development agent
  2. Choose a suitable AI-powered tool like Gemini
  3. Integrate the tool with your existing data platform
  4. Evolve the tool from a simple SQL generator to a full co-pilot
  5. Leverage the tool's advanced natural language processing and large context window handling
💡 Gemini's advanced ability to handle large context windows is a game-changer for developers working with large code bases or doing deep data analysis.

Related Reads

Up next
Anyone Can Build a Startup Now (Almost)
Jake Claver
Watch →