How to apply for Technical Program Manager, Research, FAR.Research

About FAR.AI

FAR.AI is an AI safety research nonprofit focused on addressing risks from advanced AI systems. It works on technical safety domains including model deception, interpretability, and evaluation. The role is remote in the US and centers on research program management rather than software development.

About the role

This technical program manager role oversees AI safety research programs and infrastructure from conception through delivery. You will scope plans and milestones, manage multi-team work, prioritize competing demands with research teams, build compute and resource roadmaps, and package research findings for frontier labs and other stakeholders. The work directly supports research on model deception, interpretability, evaluation, and safety.

A typical day

A typical day may involve coordinating with research teams on program milestones, reviewing compute and resource roadmaps, and resolving bottlenecks across multi-team work. You might also spend time packaging research findings or methods for frontier labs and other stakeholders. The exact daily routine is not specified in the job details, so ask about team structure and typical workflows during interviews.

Who FAR.AI is looking for

• 4+ years building technical programs from scratch, including scoping plans and milestones and managing multi-team delivery.
• Strong understanding of ML/AI concepts and research workflows, without needing to write code.
• Experience prioritizing competing research demands and focusing teams on the highest-impact work.
• Track record of building compute or resource roadmaps, identifying bottlenecks, and developing external partnerships.
• Able to package and transfer research findings, methods, and tools to frontier labs and other stakeholders.

Tips for this application

• In your resume, show specific technical programs you built from scratch, including the milestones you set and how you managed multi-team work.
• Name the AI safety domains from the job description (model deception, interpretability, evaluation, safety) and connect your experience to at least one of them.
• Describe a time you prioritized competing research demands and explain how you decided which work was highest-impact.
• Include an example of a compute or resource roadmap you built, including bottlenecks you identified and external partnerships you developed.
• Explain how you packaged research findings, methods, or tools for external stakeholders such as frontier labs, and what the outcome was.

What to cover in your cover letter

['Your experience taking complex technical programs from idea to delivery, with concrete examples of scoping plans and milestones.', 'How you prioritize competing demands with research teams and focus effort on the highest-impact work.', 'Your ability to build and maintain compute and resource roadmaps, including identifying bottlenecks and developing external partnerships.', 'Your understanding of ML/AI concepts and your experience transferring research findings, methods, and tools to frontier labs or other stakeholders.']

Research before applying

• Read FAR.AI's published research or public materials on model deception, interpretability, evaluation, and safety to understand their technical focus.
• Look into how FAR.AI describes its work with frontier labs and what transferring research findings to stakeholders means in practice.
• Check FAR.AI's public statements on compute and resource needs for AI safety research to understand the infrastructure side of the role.
• Find any public information on FAR.AI's team structure and how research programs are currently managed, or prepare to ask about this in interviews.

Likely interview topics

Based on the job description, expect questions about:

• How you would scope a new AI safety research program from idea to delivery, including milestones and multi-team coordination.
• How you prioritize competing research demands across model deception, interpretability, evaluation, and safety domains.
• How you build and maintain compute and resource roadmaps, and how you identify and resolve bottlenecks.
• How you package and transfer research findings, methods, and tools to frontier labs and other stakeholders.
• Your understanding of ML/AI concepts and how you work with research teams without writing code.

Common mistakes to avoid

• Applying without showing any familiarity with AI safety research areas such as model deception, interpretability, or evaluation.
• Describing program management experience only in general terms without concrete examples of scoping plans, milestones, and multi-team delivery.
• Claiming deep ML/AI knowledge without being able to explain relevant concepts, since the role requires strong ML/AI understanding even though coding is not required.