People at PromptEval

What's great about building PromptEval is that every decision actually matters. There's no committee, no approval chain, no six-week roadmap meeting to decide whether a feature ships. If something is worth building, it gets built. If it's not working, it gets cut. That kind of direct feedback loop is rare and I don't take it for granted. The problem we're solving is genuinely interesting. Prompt engineering sits at this weird intersection of linguistics, software engineering, and systems thinking. Every prompt is a small spec document for a non-deterministic system. Figuring out how to evaluate that rigorously is a hard problem, and working on hard problems every day makes the work feel meaningful. The pace is fast but sustainable. We ship constantly, but we also think carefully. I care a lot about the quality of what we build, not in a perfectionist way that blocks progress, but in a "this should actually work well for the person using it" way. That standard runs through everything, from the scoring logic to the copy on the dashboard. And honestly? Seeing someone share their score on Reddit, or getting a message from a developer saying the Slug API changed how their team manages prompts never gets old. Building something people find genuinely useful is still the best part of this job.