Perception Engineer, On-device ML

Teach Tuck to recognise 214 household objects without a single frame leaving the home. On-device models, tight budgets, real kitchens.

Apply for this role

What you’ll do

Train and ship the models that tell a mug from a vase and a sock from a toy. Everything runs on Tuck’s own chip, so you’ll care about milliseconds and milliwatts.

Build the tools that let families teach new objects by pointing at them once.

What you’ll bring

Experience shipping vision models to edge hardware. You like messy data and you’re suspicious of benchmarks that were shot in a lab.

How we hire

An intro call, a paired session on a real bug, and a conversation with the founders. We don’t do whiteboard puzzles.

Sound like you?

Send a CV or a portfolio, and a few lines on what you’d want to build first.

Apply by email

All roles