Evals Software Engineer (Infrastructure focus)

Apollo Research is seeking a Senior Software Engineer to lead their infrastructure initiatives for AI evaluation systems. This role combines sophisticated backend engineering with cutting-edge AI safety research, focusing on building secure and scalable systems for evaluating frontier AI models. The position offers an opportunity to work with a distinguished team of researchers and engineers in central London, contributing to critical work in AI safety and alignment.

The role involves designing and implementing infrastructure for running frontier Language Model evaluations, making key technical decisions about the stack, and building essential internal tools. You'll work closely with researchers to understand and support their needs while ensuring robust security practices across all systems. The position requires expertise in Python, Kubernetes, and AWS, with a strong focus on Infrastructure as Code and security best practices.

As part of Apollo Research's mission to understand and prevent deceptive alignment in AI systems, you'll be working on projects that directly impact the safety and reliability of frontier AI systems. The company offers a collaborative environment, working alongside team members like Mikita Balesni, Jérémy Scheurer, and others who are leading figures in AI safety research.

The position comes with excellent benefits, including competitive compensation, flexible working hours, unlimited vacation, and comprehensive relocation support through the UK Government's AI Futures Grants program. The role is based in London, sharing office space with the London Initiative for Safe AI (LISA), and offers visa sponsorship for international candidates.

This is an ideal opportunity for a senior engineer who wants to combine technical excellence with meaningful impact in AI safety, working on infrastructure that supports critical research in understanding and evaluating advanced AI systems. The role offers significant autonomy and the chance to shape the technical direction of an important player in the AI safety field.