tl;dr: research scientist / research engineer, $135-270k + equity, relocation to London

Want to test frontier models before they're released? 👁

Here's the deal: Apollo is hiring not only engineers but also researchers. OpenAI/Anthropic/Deepmind send them fresh models weeks before release, and researchers test them for scheming: deception, hidden goals, sabotage attempts. Their reports then go into the system card and influence how and when the model is released 😉

Apollo is doing very well right now: last year they officially partnered with OpenAI, CEO Marius Hobbah made it into the Time 100 AI list, and their research is in more demand than ever because reasons.

They are hiring for two roles immediately:

🎉 Research Scientist in evals — be among the first in the world to test new models, check them for scheming, examine transcripts for strange behavior, and automate the above. Requirements: strong SWE experience (ideally Python); analytics — ability to extract signals from large, noisy data; communication — ability to clearly convey results to diverse audiences. And, of course, actively understand what models can do what!

😈 Research Scientist in scheming science — build the science of why models start deceiving in the first place: study RL dynamics, create model organisms, work towards scaling laws for scheming. Keywords: experience in empirical research; experience with RL training of LLMs; you've thought and read a lot about scheming and AI cognition in general; strong math foundation (e.g., in statistical physics); SWE experience. Here it is not necessary to have everything on the list, but it's important to have solid experience in some of the areas.

Relocation to London is provided, salary range $135k-270k + equity, lunches and dinners at the office are provided, unlimited vacation and a budget for self-development. Write to Diana @yourdivna ❤️