If you care about balancing the speed, cost and quality of your product, you need to care about evals. The days of a Product Manager giving a Product Designer a PRD, and then the Product Designer chucking a Figma over the fence to engineering are gone!

A solid scope, golden dataset and rubric all help to align stakeholders and prevent regressions and drift in the product. It’s a tricky concept for people to understand, so rather than write a long how-to doc, I turned my knowledge into an interactive tool.

This tool explains why evals are important, guides you in writing your own scope, goldens and rubric, and lets you test it out using an LLM-as-a-judge. Once you’re happy, you can export your work to use in your project.

https://evals-lab.com