We're trying to apply the insights of category theory, dependent type theory, and functional programming to deep learning. How do we best equip neural nets with strong inductive biases from these fields to help them reason in a structured way? Our upcoming ICML paper gives some flavor https://arxiv.org/abs/2402.15332 ; you can also watch https://www.youtube.com/watch?v=rie-9AEhYdY&t=387s ; but there is a lot more to say.
If you are fluent in 2 or more of { category theory, Haskell (/Idris/Agda/...), deep learning }, you'll probably have a lot of fun with us!
Check out our open positions at https://jobs.gusto.com/boards/symbolica-ai-67195a74-31b4-405...
Supercede builds industry-leading risk placement and analytics software for the reinsurance industry. We're all-in on functional programming, and we're looking for more help.
Our tech stack is predominantly:
- Haskell (Yesod, Conduit, Persistent/Esqueleto, HSpec/QuickCheck, Shakespeare, Aeson)
- Nix
- PostgreSQL
- HTML, CSS, and [minimal] JavaScript
You must have demonstrable experience building things in idiomatic Haskell, and you should have a solid understanding of how web applications typically work. You should be able to communicate clearly in English, but you do not need to speak English natively. You'll have flexible working hours and you can work from anywhere, though this is a full-time position and the expectation is that of work based on a traditional 40 hour week, with 25 days of paid annual holiday in addition to your country's national holidays.
We favour asynchronous communication, and try to hire "managers of one". We don't do daily stand-ups. We don't count your hours. We don't work weekends. We support each other in working and learning, and we have 20% time where programmers are able to work on open source and/or research projects. We value colleagues who are kind, diligent, and tenacious.
Are you a good fit? Please write me a brief email introducing yourself with your résumé attached to jezen@supercede.com.