// AI engineeringDec 2024 · 7 min read · by Reco
Evals for the rest of us: a minimum-viable eval rig in 2 hours
You don't need a research team to know if your agent is getting worse. A pragmatic eval setup any founder can run from day one.
// draft
The full write-up is being edited and will publish here shortly. In the meantime, if this topic is relevant to a build you’re planning, book a roadmap call — we’ll walk through it directly.