Latest AI JobsBrowse Jobs
Thinking Machines logo

Research, RL Scaling

Thinking MachinesSan FranciscoResearch
Apply on company site
Salary
$350K - $475K USD
Job type
Full-time
Work mode
Remote
Posted
21 Aug 2026
Status
Open · checked 7d ago

About this role

The Research, RL Scaling role at Thinking Machines is a full-time position based in San Francisco, California, with remote flexibility, offering individuals the opportunity to scale reinforcement learning for frontier models. This position suits researchers who want to own the boundary between training recipes and infrastructure while advancing asynchronous reinforcement learning.

Role Overview

Thinking Machines is hiring a researcher to focus on high-training-compute, long-horizon reinforcement learning. This position involves full-stack ownership, ranging from algorithms and parallelism plans to the overall health of training runs. The work centers heavily on asynchronous reinforcement learning, requiring a deep understanding of async algorithms, systems design, and trade-offs across both machine learning and systems.

Key Responsibilities

Co-design the reinforcement learning recipe and the systems that run it, validating choices at frontier scale.
Advance asynchronous reinforcement learning algorithms.
Improve rollout generation efficiency and its integration with training, treating inference as a first-class part of the loop.
Run frontier-scale reinforcement learning end to end by bringing up new models and setups while maintaining stability.
Jointly optimize compute and training efficiency, including accelerator utilization, memory, communication, and low-precision numerics.
Conduct careful empirical science through trustworthy instrumentation, clear ablations, and scaling studies.

Requirements

Proficiency in Python and familiarity with at least one deep learning framework such as PyTorch, TensorFlow, or JAX, along with comfort in debugging distributed training.
Bachelor’s degree or equivalent experience in Computer Science, Machine Learning, Physics, Mathematics, or a related discipline.
Clear written communication skills with the ability to explain complex technical concepts.
Strong research judgment demonstrated through clean ablations, honest baselines, and clear technical writing.

Nice to Have

PhD in Computer Science, Machine Learning, Physics, Mathematics, or a related discipline, or equivalent industry research experience.
Strong grounding in reinforcement learning for large language models, including modern policy optimization methods.
Deep understanding of asynchronous reinforcement learning algorithms, design choices, and trade-offs.
Experience training large models across many accelerators and familiarity with the distributed stack.
Working knowledge of inference systems, rollout generation throughput, and cost reasoning.
Experience building or operating decoupled generation and training reinforcement learning systems at scale.
Experience with reinforcement learning on verifiable and agentic tasks, including multi-turn environments.
Experience with reinforcement learning training stability techniques for large runs.
Familiarity with low-precision training, inference numerics, quantization, and LLM serving stacks.
Contributions to open-source training or inference frameworks and scaling studies for large models.

Skills and Tools

Python
PyTorch
TensorFlow
JAX
Large Language Models (LLMs)
Distributed training stacks
Inference systems and serving stacks

Salary and Benefits

Expected annual salary range of $350,000 to $475,000 USD, depending on background, skills, and experience.
Generous health, dental, and vision benefits.
Unlimited paid time off (PTO).
Paid parental leave.
Relocation support as needed.
Visa sponsorship available.

About the Company

Thinking Machines builds artificial intelligence designed to extend human will and judgment. The organization is training frontier models using Inkling, developing Tinker to allow people to customize models, and crafting interfaces to broaden human-AI communication with the belief that a human future is worth building.

How to Apply

Interested candidates should apply directly on the official company careers page.

Skills and tech stack

PythonPyTorchTensorFlowLLM

Questions candidates ask

What is the location of the Research, RL Scaling role?

The role is based in San Francisco, California, with remote flexibility available.

What is the salary range for this position?

The expected annual salary range is $350,000 to $475,000 USD, depending on background, skills, and experience.

Does Thinking Machines offer visa sponsorship?

Yes, Thinking Machines sponsors visas for qualified candidates.

What benefits are provided?

Benefits include generous health, dental, and vision coverage, unlimited paid time off, paid parental leave, and relocation support as needed.

What are the primary technical skills required?

Candidates must be proficient in Python, familiar with at least one deep learning framework like PyTorch, TensorFlow, or JAX, and comfortable with debugging distributed training.

Ready to apply?

The application is on the official Thinking Machines careers page.

Apply on company site