🎄 Let's code and celebrate this holiday season with Advent of Haystack

Advent of Haystack

Welcome back to another year of Haystack challenges

with 10 challenges in the month of December 🎉

Complete and submit all challenges by December 31 for a chance to win gift cards, swag, and more! 🎁 Learn more in Advent of Haystack

✨🎄 Plus: Share Your Haystack Story This Holiday Season! 🎄✨

Spread the cheer and get perks by sharing your journey with Haystack. See How.

Discuss
Submit

Day 5: Elves’ Secret for Faster Development 💨

In the buzzing Tech Hub of the North Pole, Elf Tobi was at work, surrounded by glowing monitors and mugs of hot cocoa. Known for building quick and modular solutions, Elf Tobi thrived on AI challenges. Because he had a secret 🤫

Elf Tobi was a Haystack fan and had recently started using deepset Studio to make the pipeline-building process even more fun and efficient. Its drag-and-drop interface made building and editing Haystack pipelines a breeze, turning a challenging task into a fun, interactive experience. He especially appreciated the deployment and testing features, as well as the handy pipeline templates that provided a great starting point. “Deploying and testing on the Playground? That was like adding a shiny star on top of the Christmas tree!” Elf Tobi thought about deepset Studio 🎄

For this challenge, you’ll join Elf Tobi to create a RAG pipeline in deepset Studio. Here are the steps to complete the challenge:

  • Use sample files for “AI” in deepset studio to populate the Elves’ AI Library
  • Explore pipeline templates and use the “RAG Question Answering GPT-4o” template (or any other template) to save time 😉
  • (Optional) Replace the models in the pipeline, use different providers or open source ones.
  • Build, save, and deploy your indexing and query pipeline (Deployment might take a bit of time, so feel free to revisit earlier Advent of Haystack challenges while you wait ☃️)
  • Use the Playground to test your pipeline and submit the pipeline’s output for the query “What’s Compound AI?”

🎯 Helpful Links:

  • Learn more about deepset Studio features in our announcement post
  • If you have any questions, feel free to ask in the “#deepset-studio” channel on Discord

💡 Some Hints:

⭐️ Bonus Task

  • Extend your pipeline’s capabilities and build a RAG pipeline with a fallback to the web. This updated pipeline should query the Elves’ AI Library first, and if no answer can be found, fetch it from the web.
  • As the fallback, you can either just search for the results on the Web or use these Web results to generate the complete answer.
  • Get your free API key from SearchAPI or Serper and add the key to your connections/secrets to use web search components in your pipeline.
  • Get inspiration from Tutorial: Building Fallbacks to Websearch with Conditional Routing to understand how to design the pipeline with a fallback mechanism.

💙 GET STARTED: Sign up for deepset Studio (free and open to everyone) and get access within 24 hours