Co-design model capabilities and define areas for improvement with researchers.
Review academic and industry literature to identify opportunities for improvement.
Run experiments and benchmarks using the internal simulation stack and open-source tooling.
Speak with individual users and enterprises to identify model and agent underperformance, then turn insights into evaluation and training data.
Summarised from the posting by AI and can be wrong. The full description is the source.
Requirements
Skills
Deep understanding of evaluations, reinforcement learning, and how LLMs and agents work
Ability to read papers and keep up with state-of-the-art literature
Highly proficient with coding agents such as Codex or Claude Code
Other requirements
Extreme agency and a founder mentality; resourceful and scrappy
Benefits
Health insurance
Health benefits
Dental & vision
Dental and vision benefits
Paid time off
Flexible PTO
Equity
Meaningful early-stage grant
Bonus
Sizable performance bonus tied to project and company milestones
Full job description
From Abundant’s careers site
ABOUT ABUNDANT
Abundant is an applied research lab focused on scaling reinforcement learning for safe and reliable agentic capabilities. We are an extremely talent-dense team of researchers, roboticists, founders, and operators whose work includes the Waymo Driver.
THE ROLE
Research PM (RPM) is a role unique to Abundant. RPMs are PMs of the model, unlike traditional PMs who own an application or a feature. RPMs design and implement model capabilities, either as part of an AI lab, data lab, or other research organization. This is a founding-level role with outsized ownership.
WHAT YOU’LL DO
Co-design capabilities of the model: define areas of improvement for the model, along with the Researcher.
Show full descriptionShow less
Perform literature review: headroom analysis, identifying recent work in academic and industry circles.
Run experiments and benchmarks: using our internal simulation stack and our open source tooling such as Oddish to define and hillclimb on the vertical.
Become the voice of the model user: Talk to individual users and/or enterprises, in order to deeply understand where models and agents are underperforming, turning these insights into evaluation and training data.
Work on data pipelines: internal tooling to support research operations.
WHO YOU ARE
Technically fluent: Deep understanding of evaluations, RL, and how LLMs and Agents work.
Research-oriented — Able to read papers and keep up with SOTA literature.
Extreme owner — You have extreme agency and a founder mentality; you are resourceful and scrappy.
Creative problem solver — You leverage creativity and research skills to design projects from the ground up.
Must be highly proficient with coding agents such as Codex or Claude Code.
COMPENSATION
Base Salary
$150,000 - $210,000
Cash Bonus
Sizable performance bonus tied to project and company milestones