Task Development Engineer

• In this role, you'll develop challenging, novel evaluation tasks for frontier AI models that remain difficult as model capabilities grow.
• Conduct quality assurance on existing tasks to verify solvability and ensure models receive only necessary information.
• Baseline and score task completions using your domain expertise to validate evaluation quality.
• Improve task development infrastructure by identifying inefficient workflows and implementing enhancements.
• Stay current with evaluation methodologies and frontier AI capabilities to inform effective task design.

• In this role, you'll develop challenging, novel evaluation tasks for frontier AI models that remain difficult as model capabilities grow.
• Conduct quality assurance on existing tasks to verify solvability and ensure models receive only necessary information.
• Baseline and score task completions using your domain expertise to validate evaluation quality.
• Improve task development infrastructure by identifying inefficient workflows and implementing enhancements.
• Stay current with evaluation methodologies and frontier AI capabilities to inform effective task design.

Model Evaluation and Threat Research (formerly Alignment Research Center, Evaluations) is a project focused on evaluating the capabilities and alignment of advanced ML models.