Infrastructure

Remote, Europe

Full-time

Inference engineer

Make long thinking cheap and fast: batching, caching and kernels for 30-minute requests.

Apply for this role

€140k–€200k + equity

Our requests can run for half an hour. Serving them well is an unsolved problem, and it is yours.

What you’ll do

  • Write and tune GPU kernels

  • Design scheduling for very long requests

  • Cut first-token latency further

You have shipped high-performance systems and like measuring before optimising.

Start here

Ask it somethinghard.

Kerrfield
An independent research lab building models that reason before they answer.
© 2026 Kerrfield Research Ltd. · London · MontréalAll systems normal

Create a free website with Framer, the website builder loved by startups, designers and agencies.