Software Engineer
Polymath · Remote · San Francisco ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- Fully remote
- AI agent simulations
- Backed by Y Combinator
- Flat organization
- Research-focused
Join Polymath, an applied research lab backed by Y Combinator and Base10, to advance the capabilities of safe, autonomous AI agents. You will engineer the simulation environments, complex tasks, and verification systems that allow frontier models to learn and operate in realistic settings.
The team works closely with leading model labs to push the boundaries of agent performance. You will solve open-ended technical problems, building robust infrastructure from first principles to support rigorous training and evaluation.
What you'll do
- Build diverse, high-fidelity simulation environments for realistic agent testing
- Design complex tasks requiring long-horizon reasoning and tool usage
- Develop robust verifiers to reliably measure agent performance
- Improve infrastructure and tooling for running and debugging environments
- Collaborate with researchers to identify failure modes and create new benchmarks
What you'll need
- Strong engineering fundamentals
- Ability to build from first principles
- Experience solving open-ended technical problems
- High agency and bias toward shipping results
- Commitment to building robust, high-quality systems
How you'll work
- Flat organizational structure
- Self-motivated work environment
- Focus on scientific progress and learning