Blog

Notes on fair hiring

Thinking on evidence-based screening, the decisions buried in a pile of applications, and how to give every candidate a fair first read.

HiringJuly 15, 20266 min read

The first ten problem

Two hundred people apply for one role. A small team reads maybe forty of them, from the top. The strongest person can be applicant one hundred and thirty, and never get a fair look.

Elena Marsh · Co-founder and CEO

Read post

Latest

June 24, 2026

Fairness

Read applications against the role, not each other

Most screening ranks candidates against each other, which quietly rewards whoever looks most like the last person you hired. Reading each application against the role is a different, fairer thing.

Daniel Okafor

7 min

May 28, 2026

AI-native

Fair screening has to be AI-native

You cannot make a keyword filter fair by adding more keywords. Reading an application against a role, with reasons, is the whole job, and it is only possible with modern language models.

Daniel Okafor

6 min