Braintrust

We’re thrilled to announce that we've raised a $36M Series A led by Martin Casado at Andreessen Horowitz to advance the future of AI software engineering, bringing our total funding to $45 million. Through our work with top AI engineering and product teams from Notion, Stripe, Vercel, Airtable, Instacart, Zapier, Coda, The Browser Company, and many others, we’ve had a front-row seat to what it takes to build world-class AI products. Along the way, we’ve learned a few key lessons: - Crafting effective prompts requires active iteration. - Evaluations are crucial for systematically improving quality over time. - Production logs provide a vital feedback loop, generating new data points that drive better evaluations. Evals are just the first step to building AI apps. That’s why we’re also excited to introduce functions, the flexible primitive for creating prompts, tools, and scorers that sync between your codebase and the Braintrust UI.

Happy valentine's day from all of us at Braintrust.

Ready for an inside look into building Zapier Agents? Join us tomorrow, January 30 at 9AM PT to learn from Braintrust CEO Ankur Goyal and Vitor Balocco of Zapier. We’ll discuss agent evals, scoring techniques, and more. Grab your spot in the comments below

How do you write better scoring functions for your evals? Loom shared their four-step process for evaluating one of their most-loved AI features: auto-generated video titles. Link to the full post is in the comments 👇 Huge thank you to Matt Granmoe for sharing these insights! #ai #product #evals

Don’t miss our webinar one week from today, on January 30th at 9AM PT. Our CEO Ankur Goyal will share practical strategies for evaluating agents in production, from picking success metrics to choosing scoring functions. Plus, Vitor Balocco from Zapier will join us to share how Zapier Agents relaunched bigger and better this week with the help of evals. Reserve your seat for next week’s webinar in the comments below.

Join our CEO Ankur Goyal on Thursday, January 30th at 9AM PT for a live session on how to evaluate AI agents effectively. We'll break down the basics of getting started with evaluations, choosing or writing scoring functions, and applying these concepts with practical examples from Anthropic's recent guide to building effective agents. We'll also be joined by Vitor Balocco from Zapier, who will share how his team has integrated evaluations into their CI/CD pipeline for their agentic features. This is a great opportunity to learn actionable steps to improve your own AI projects and hear directly from experts applying these methods in production. Registration link in the comments 👇

The Braintrust team is growing! We recently welcomed Olmo Maldonado, Mike Deeks & Jason Miller to the team. We enjoyed sundaes at Ghirardelli Square and also celebrated Jack Gardner & Manu Goyal's birthdays. If you're interested in working with us, check out our careers page: www.braintrust.dev/careers

It seems like every week, another AI provider launches a new, state-of-the-art model. They publish benchmarks that convey what it’s good at, but those benchmarks rarely map to real-life use cases. We wrote about the exact steps you should take to decide if the new model is worth deploying in your app. Link is in the comments. PS: Gemini 2.0 Flash is now available on Braintrust.

👻

Happy halloween! There's nothing spookier than building an AI app without evals.