Software Engineer, Benchmarking

• Run and maintain benchmarking infrastructure to evaluate frontier AI models.
• Implement AI benchmarks within the evaluation infrastructure to expand tracked capabilities.
• Develop existing benchmarks for quick evaluation of new model releases.
• Contribute to developing brand new benchmarks and pitch novel evaluation ideas.

This role enables deeper understanding of AI development by evaluating frontier AI models, contributing to safer and more informed AI progress.

• Experience with software engineering and benchmarking infrastructure.
• Familiarity with AI models and evaluation methodologies.
• Ability to collaborate with researchers, analysts, and engineers.

Compensation and perks not specified.

Epoch AI is a research institute focused on understanding and forecasting the trajectory of AI development, particularly in the context of AI safety and governance.