Train and serve large models in Australia
You can do this work without moving overseas, and MC-3 and our own cloud will need many more engineers and researchers.
Why Maincode
Work anywhere in the stack, from kernels to Matilda.
Each layer runs on machines we own in Melbourne, from the kernels on the AMD MI355X to the way Matilda behaves when the answer isn’t clear.
What we value
Six things we look for.
Ownership
You don’t wait to be told what to build.
Gradient
You learn new things quickly.
Range
You go deep in whatever field the problem needs.
Speed
You ship early and change course on what you learn.
Taste
You care about the details that others skip.
Humility
You care more about the work than about who gets the credit.
Open roles
Roles open now, by team.
If none of these fits, write to us about the work you want to do.
Forward Deployed Engineering 1
Data 1
Residency 1
Talent 2
Maincode Ventures 1
Research residency
A paid residency
for late-stage PhD students.
Residents spend 3 to 6 months on their own question, with our GPU cluster, our stack and our team, and strong work can ship inside Matilda.