Announcing the $100,000 BigQuery AI Hackathon

Google Cloud Community Articles & News Community News & Announcements
1

TL;DR: BigQuery AI Hackathon is here - register on Kaggle.

Structured data only tells a fraction of the story. The rest - often the most insightful part - is locked away in messy, unstructured formats like text string columns, PDFs, images, documents, and audio files. Traditionally, analyzing it required wrestling with complex data pipelines, keeping it walled off from your primary analytics workflow.

It’s time to tear down that wall. Imagine analyzing unstructured text, images, audio, and documents right alongside your tabular data, using the same SQL and Python skills you already know and love.

Build what’s next at the BigQuery AI Hackathon

We’re challenging you to build what’s next in AI and compete for your share of the $100,000 prize pool.

You’ll get hands-on experience with BigQuery’s newest features that bring AI directly to your data. SQL users will find these capabilities feel like a natural extension of their existing workflow, while Python practitioners can use BigQuery DataFrames to work using a familiar, pandas-like API. The goal is simple: build powerful, scalable AI solutions right where your data lives.

Choose your challenge

  • Generative AI: Go beyond simple queries. Use BigQuery’s LLM functions and Gemini models to generate content, extract structured data, or even forecast future trends - no training required. For example, you can use AI.GENERATE_TABLE to pull key details like issue type, product names, and sentiment from messy customer support logs into a new table with a defined schema - no complex parsing required. Or you can use AI.FORECAST (Python) for zero-shot time series predictions.

  • Vector Search: Move beyond keyword matching and find information based on what it means. With BigQuery, you can generate embeddings (Python) and use vector search (Python) to find semantically similar items. Think of building a tool to surface past patents that are conceptually similar to a new idea or finding the perfect product substitute based on its photo or description - not just its name.

  • Multimodal: Break down your data silos. Use BigQuery’s ObjectRef to analyze multimodal data (like images, audio, or documents) alongside your structured data in a single SQL or Python query. For example, you could improve property value predictions by analyzing street-view imagery in the same query as structured data like square footage and number of bedrooms

More than just bragging rights

We’re putting up a $100,000 prize pool to reward the most innovative projects.

  • For each track (Generative AI, Vector Search, and Multimodal), we’re awarding:

    • 1st Place: $15,000

    • 2nd place: $9,000

    • 3rd place: $6,000

  • Honorable mentions (two awards): $5,000 each

Beyond the cash prizes, this is an opportunity to build a public portfolio, create high-quality code samples and collaborate with the community, and gain experience with the latest Google Cloud technologies.

Get ready to build!

Ready to get started? Here’s how:

  1. Register now: Head over to the official competition website to secure your spot.

  2. Mark Your Calendar: The hackathon kicks off on August 11 and runs for 6 weeks, giving you plenty of time to build something amazing.

  3. Join the Community: Don’t build alone! Join the discussion on the Kaggle forums to connect with other participants and brainstorm ideas.

We can’t wait to see what incredible solutions you build. Good luck!

2 Likes