Return to jobs
Join Apollo Research as an Evaluations Engineer for a unique opportunity to work with frontier AI companies and test new models before anyone else.
Overview
Salary
£100000 - £200000
Location
London, UK
Expires
Expires at anytime
Organisation summary
Apollo Research is an innovative company at the forefront of AI technology. They work tirelessly to ensure control over AI and mitigate risks of deception or misuse. Their comprehensive research extends from detection to science to mitigation measures. Working closely with AI frontier companies, they provide invaluable testing and feedback before model deployment. Apollo is dedicated to truth-seeking, being goal-oriented and hence promotes a very friendly, constructive work culture.
Role Summary
- Run and own 'evaluation campaigns' involving pre-deployment testing of models.
- Develop the evaluation infrastructure and automate the pipeline.
- Work collaboratively with leading AI labs including OpenAI, Anthropic, and Google DeepMind.
Role Requirements
- Proficiency in Python, process optimization, data analysis, pattern recognition and strong communication skills.
- Experience with Inspect evals framework is a plus, though not required.
- Self-taught candidates are welcome; no formal background or industry experience is required.
Application Process Details
- Complete an application form and submit your CV; cover letter not necessary.
- Submit links to relevant work samples if you wish.
- The multi-stage interview process consists of a screening interview, a take-home test, three technical interviews, and a final interview with the CEO.
ABOUT APOLLO RESEARCH The rapid rise in AI capabilities offer tremendous opportunities, but also present significant risks. At Apollo Research, we're primarily concerned with risks from Loss of Control, i.e. risks coming from the model itself rather than e.g. humans misusing the AI. We're particularly concerned with deceptive alignment / scheming, a phenomenon where a model appears to be aligned but is, in fact, misaligned and capable of evading human oversight. We work on the detection of scheming (e.g., building evaluations), the science of scheming (e.g., model organisms), and scheming mitigations (e.g., anti-scheming and control). We closely work with multiple frontier AI companies, e.g. to test their models before deployment or collaborate on scheming mitigations. At Apollo, we aim for a culture that emphasizes truth-seeking, being goal-oriented, giving and receiving constructive feedback, and being friendly and helpful.
THE ROLE As an Evaluations Engineer, you will run and own 'evaluation campaigns' (pre-deployment testing for unreleased frontier models), build out our evaluation infrastructure, and automate the evals pipeline. You will get to work with frontier labs like OpenAI, Anthropic, and Google DeepMind and be amongst the first to interact with new models before anyone else.
JOB REQUIREMENTS We don't require a formal background or industry experience and welcome self-taught candidates. Key skills include software engineering with Python, process optimization, data analysis & pattern recognition, and strong writing and communication abilities. Experience with Inspect evals framework is a bonus.
BENEFITS Flexible work hours and schedule, unlimited vacation and sick leave, lunch, dinner, and snacks provided on workdays, paid work trips, and a yearly $1,000 (USD) professional development budget.
HOW TO APPLY Please complete the application form with your CV. The provision of a cover letter is not necessary. Please also feel free to share links to relevant work samples. Our multi-stage interview process includes a screening interview, a take-home test, 3 technical interviews, and a final interview with the CEO.
