Build the next architecture.
We are a small team working on hard, open problems. If you like measuring things carefully and shipping research that works, we would like to hear from you.
Where you could fit
Research Scientist — Efficient Architectures
Memory, attention alternatives and certified retrieval. You design experiments and read the curves honestly.
ML Systems Engineer — GPU kernels
Fused kernels and level-packed memory layouts so tree reads show up in wall-clock time.
Research Engineer — Training & Data
Taxonomy-driven datasets, ternary training and quantization-aware pipelines.
Research Intern
Own a small, well-defined experiment from hypothesis to write-up.
Open application
Do not see your role? Tell us what you would work on.
What to expect
Real problems
You will work directly on the architecture, not on wrappers around someone else’s model.
Open by default
Results, including failed ones, go into a public research log.
Small team
Direct access to the founder, fast decisions and wide ownership.