Latest AI JobsBrowse Jobs
P

Member of Technical Staff - Research Engineer, Post-training

Preference-ModelSan FranciscoEngineering
Apply on company site
Salary
$200K – $350K
Job type
Full-time
Experience
Senior
Posted
11 Sept 2026
Status
Open · checked 7d ago

About this role

Based in San Francisco, this full-time Member of Technical Staff position focuses on post-training research engineering at Preference-Model and suits individuals with practical background in running large language model pipelines. The organization seeks a professional to help advance reinforcement learning methodologies for advanced intelligence systems.

Role Overview

Preference-Model is hiring a Member of Technical Staff in San Francisco to serve as a Research Engineer focusing on post-training. This position combines research and engineering responsibilities to investigate self-directed learning boundaries and implement new approaches for reinforcement learning.

Key Responsibilities

Train and evaluate models using proprietary reinforcement learning environments to check data quality, identify gaps in tasks, and connect environment design with model capabilities.
Build and improve reinforcement learning training infrastructure, ranging from training abstractions to managing distributed experiments with frameworks like Verl and OpenRLHF, while scaling systems for complex research.
Design, build, and test training environments, evaluation methods, and frameworks for reinforcement learning agents.
Monitor and enhance training runs completely from data loading to reward calculation to increase throughput and speed up research iterations.

Requirements

Background executing complete post-training pipelines for large language models with at least 7 billion parameters.
Working knowledge of Python alongside PyTorch or JAX.
Practical background using at least one modern reinforcement learning training framework.
Competence in building and running machine learning infrastructure at scale.

Nice to Have

Background assessing model outputs and creating reward or evaluation signals.
Awareness of current post-training research literature and the ability to turn academic papers into functional code.
Opinions regarding code structure for reinforcement learning training to ensure reproducibility and rapid iteration.
Ability to combine research exploration with engineering rigor.
Systems design and communication skills.

Skills and Tools

Python
PyTorch
JAX
Verl
OpenRLHF

Salary and Benefits

Compensation ranging from $200K to $350K with equity offers
Health, vision, and dental benefits
401K match
Daily onsite lunch and weekly snack orders
Visa sponsorship and relocation support

About the Company

Preference-Model operates as a superintelligence data research company building reinforcement learning environments to train capable and aligned systems. The organization secured $16M in seed funding led by a16z and builds upon prior experience developing data infrastructure and datasets for Claude at Anthropic.

How to Apply

Interested candidates should visit the official company careers page to submit an application.

Skills and tech stack

PythonPyTorchJAXVerlOpenRLHFLLMMachine Learning

Questions candidates ask

What is the location of the position?

The role is located in San Francisco.

What salary and financial compensation does the role offer?

The position offers a salary between $200K and $350K along with equity offers.

Is a PhD required to apply for this role?

No, candidates do not need a PhD or extensive publications to be considered.

Are visa sponsorship and relocation support available?

Yes, visa sponsorship and relocation support are available for this position.

What specific frameworks are mentioned for training infrastructure?

The posting lists frameworks like Verl and OpenRLHF.

Ready to apply?

The application is on the official Preference-Model careers page.

Apply on company site