Software Engineer

Polymath; San Francisco (CA); On-site; USD 100,000 - 150,000; Part time

Job summary

Polymath is seeking Software Engineers in San Francisco to build simulation environments that challenge advanced AI models. This contract role offers a pathway to full-time employment and involves creating realistic tasks for agents and improving measurement systems. Ideal candidates will have strong engineering principles, high agency, and a commitment to delivering robust solutions. Join a motivated team dedicated to pushing the boundaries of safe AI development.

Qualifications

• Strong engineering fundamentals are required.
• Enjoy building from first principles and solving complex technical problems.
• High agency and a strong bias toward shipping are essential.

Responsibilities

• Build diverse, high-fidelity environments to test agents.
• Design tasks requiring long-horizon reasoning and tool use.
• Develop verifiers to reliably measure agent performance.
• Improve infrastructure to debug and enhance environments.
• Work with the research team to identify and address failure modes.

Skills

Strong engineering fundamentals

Building from first principles

Solving open-ended technical problems

High agency

Strong bias toward shipping

Building robust systems

Job description

About Polymath

Polymathisan applied research lab focused on advancing long-horizon agent capabilities through reinforcement learning. We design and scale simulation environments where agents learn to operate safely and autonomously. We work with the world’s leading model labs to push the frontier of agent capabilities. Polymath is backed by Base10, Founders Future, Y Combinator, and other incredible investors & angels. We've raised an $8M seed, and are actively growing out the team.

About the role

We’re hiring Software Engineers to build the simulation environments, tasks, and verifiers that challenge frontier models. You’ll help create the training and evaluation grounds that make it possible to measure and improve autonomous agents on realistic, challenging work. This is a contract-based role with the opportunity to transition into a full-time position.

Examples of projects you could work on include:

• Building diverse, high-fidelity environments that test agents in realistic settings

• Designing complex tasks that require long-horizon reasoning and tool use

• Developing robust verifiers that reliably measure agent performance

• Improving infrastructure and tooling to run, debug, and improve environments

• Working closely with the research team to identify failure modes and turn them into new tasks and benchmarks

You’ll be a good fit if you:

• Have strong engineering fundamentals

• Enjoy building from first principles and solving open-ended technical problems

• Have high agency and a strong bias toward shipping

• Have a high quality bar and care about building robust systems

Culture:

• Polymath is a team of researchers, engineers, and operators focused on advancing the frontier of safe, superintelligent AI agents.

• We have a flat organizational structure. We believe that people do their best work when they’re self-motivated and driven by a desire to learn, contribute to the team’s goals, and advance scientific progress.

• We’re looking for folks who ship fast, set high standards for themselves, and are great team players.