Lead AI Research Programs: Post-Training & Evaluation

AfterQuery; San Francisco (CA); On-site; USD 120,000 - 180,000; Full time

Benefits offered by this job

Health Insurance

401(k) with Employer Match

Daily UberEats Stipend

Monthly Wellness Stipend

Job summary

AfterQuery in San Francisco is seeking a program manager to move ambitious research programs forward. You will own execution across post-training, enterprise post-training, fellowships, evaluation releases, and partner research, turning open-ended goals into concrete plans.

You will align researchers, engineers, SPLs and external partners, surface blockers, and decide on priorities. This is not traditional PM; you must understand technical work well enough to reason about experiments and

Qualifications

• 2-4 years project/program management experience.
• Experience leading complex technical programs across research, machine learning, engineering, or data organizations.
• Working knowledge of the LLM development lifecycle, including data creation, SFT, RL-based post-training, and evaluation.
• Strong communication and an ability to turn messy technical information into clear decisions and next steps.

Responsibilities

• Own planning and execution across post-training, enterprise post-training, fellowship, and public evaluation programs.
• Translate ambiguous research objectives into milestones, owners, dependencies, success criteria, and decision points.
• Maintain a trustworthy view of program health, including experiment progress, risks, blockers, partner dependencies, and upcoming releases.

Skills

Project management

Cross-functional leadership

ML lifecycle knowledge

Communication skills

Job description

AfterQuery in San Francisco is seeking a program manager to move ambitious research programs forward. You will own execution across post-training, enterprise post-training, fellowships, evaluation releases, and partner research, turning open-ended goals into concrete plans.

You will align researchers, engineers, SPLs and external partners, surface blockers, and decide on priorities. This is not traditional PM; you must understand technical work well enough to reason about experiments and