Secure Program Synthesis Fellowship

• Conduct research in a small team on secure program synthesis, addressing specification, validation, and adversarial robustness of AI-generated code.
• Explore focus areas including specification elicitation, validation workflows, or robustness evaluation.
• Produce research artifacts and workshop papers documenting findings throughout the program.
• Present milestones and prepare work for demo day presentations and potential conference publication.

This fellowship advances the safety and reliability of AI-generated code, reducing risks from specification failures and adversarial attacks, thereby contributing to the broader goal of safe AI development.

• Interest in AI safety and program synthesis research.
• Ability to work collaboratively in a small team.
• Strong analytical and problem-solving skills.
• Background in computer science or related field (preferred).

Fellowship program with expert mentorship, project management support, and opportunities for conference publication. No salary mentioned.

Apart Research is an organization focused on AI safety research, conducting programs and fellowships to address critical challenges in AI alignment and robustness.