Computer Vision Interview Questions, issue 20, Jan 21, 2026
The Low-Contrast Bias Trap
Senior Computer Vision Engineer interview at OpenAI, and the interviewer asks:
“Our production FaceID model has a 12% higher error rate on darker skin tones. We audited the training data and it is perfectly balanced (50/50 split). We retrained from scratch. The error persists. Why?”
Don’t say: “You need to oversample the underrepresented class,”
Why perfectly balanced face datasets still fail - and how sensor physics silently breaks feature learning.
The full answer, with the mechanism and the arithmetic, is for paid subscribers on Substack.