DAGER: Deep Age, Gender and Emotion Recognition
Needs reviewThis page has no source text on the previous site — the copy below is drafted from the résumé and needs a pass before launch.
Problem
Commercial identity systems needed age, gender, and emotion estimation accurate enough to make decisions on — and it had to run on the device, not in a datacenter.
Approach
A multi-task convolutional network sharing one backbone across all attributes, so the marginal cost of each additional prediction stayed near zero. Tuned end to end for on-device inference rather than benchmark accuracy.
Result
Under 1% false accept and reject rates in commercial deployments, including in-car identification. Powered the full Sighthound facial analytics product line.