Careers

Work on questions worth asking.

We are building a research organization around hard, open problems in machine intelligence. We are interested in hearing from people who want to work on them carefully.

Open roles

Open roles

No open roles listed at the moment.

We will post positions here when they exist. If our research overlaps with yours, we would still like to hear from you — tell us what you work on and which of our questions interests you.

Express interesthello@theoryandsignal.com

Who we hope to hear from

Many kinds of expertise, one shared standard.

Research on intelligence draws on more disciplines than any one person can hold. These are the profiles we expect to matter. They describe interests, not open positions.

  • 01

    Research scientists

    People who formulate sharp questions about learning, reasoning, or perception — and design the experiments that answer them.

  • 02

    Machine learning researchers

    Specialists in optimization, representation learning, or evaluation who care as much about why a method works as whether it does.

  • 03

    Research engineers

    Engineers who turn ideas into reliable experimental infrastructure: training pipelines, evaluation harnesses, and reproducible tooling.

  • 04

    Software engineers

    Builders of the systems that research depends on — data platforms, distributed compute, and well-designed internal tools.

  • 05

    Computational scientists

    Physicists, biologists, chemists, and mathematicians who see where learned models could change how their field works.

  • 06

    Independent collaborators

    Researchers at other institutions, or working independently, who are pursuing related questions and want to compare notes.

How we intend to work

What you can expect from us.

Questions first
Projects begin with a clearly stated question and a view of what evidence would answer it.
Time to think
Research rarely proceeds on a schedule. We plan for depth over volume.
Open exchange
We aim to publish, share methods, and engage with the wider research community.
Shared responsibility
Everyone considers reliability and consequences — not only a designated team.