What it is
A local screening workbench for a folder of CVs, built on TypeSafe's Jev model. It is for anyone hiring who wants a sortable shortlist with one verdict per candidate.
How it uses Jev
It asks Jev a small fixed set of typed questions about each CV, then scores the answers with plain arithmetic you can read. The role being hired for is a policy you edit in the app; changing a weight, a cap or a question re-scores every stored candidate in about 20 ms and costs nothing.
Primitives:choicescorenoul
Technique worth stealing
Keep the model's judgments separate from the arithmetic, so policy edits re-score stored candidates for free.
Try it
See cv-screen/README.md for how to run it.
Evidence
Each line is one question put to Jev about the README. ≥ 0.60 reads as yes, ≤ 0.40 as no; in between Jev is not making a call.
- Jev-centricyes0.94
- Shows a System One patternyes0.78
- Handles uncertaintyno0.04
- Measuredno0.09
- Runnableyes0.76
- Worth recommendingyes0.61
- Model replicano0.08
- Problem scopescore on a 0–2 scale1.16
- About Jevyes0.98
Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 20 Sept 2026.