Full-Time

Founding Engineer

Proximal

Proximal

1-10 employees

AI training environments for coding agents

Compensation Overview

$200k - $300k/yr

+ with a generous equity grant.

H1B Sponsorship Available

San Francisco, CA, USA

Category
Software Engineering (2)
,
Requirements
  • Strong systems engineer. 3+ years designing, building, or scaling distributed or infra-heavy systems.
  • Previous industry experience. Previous work at dev tooling, infra, or AI system companies (e.g. Modal, Cursor, Anthropic).
  • Green GitHub. Consistent commit history and well-maintained repos showing long-term engagement.
  • Builder track record. Has spent months refining and maintaining a personal or collaborative project — ideally something others actively use.
  • Bonus: Advanced tools. Uses or contributes to dev tooling, or prefers advanced languages and custom setups (Rust, SolidJS, or Vim).
  • Bonus: Technical writing. Has a personal site or blog with posts on engineering concepts or systems design.
Responsibilities
  • Design and build distributed systems that run thousands of concurrent simulations with fast, memory-efficient snapshotting and recovery.
  • Develop abstractions that make environment building easier — small-loop agents, infra mocking, SDK stubbing, and related tools.
  • Prototype and productionize new environment types; build systems to generate training data automatically at scale.
  • Build and evaluate AI agents that analyze large production codebases and generate realistic pull requests.
  • Experiment with AI copilots that improve speed and quality across environment engineering workflows.
  • Work hands-on across infrastructure and agents, running experiments end to end and learning fast.
  • Help define and grow the technical foundation that powers the next generation of coding models.

Proximal builds simulation-based training environments that enable AI models to learn software engineering tasks through experience, not static data. By creating realistic, complex simulations of real-world engineering challenges, Proximal helps frontier labs and enterprises train next-generation coding agents capable of superhuman performance. The team—alumni of OpenAI, Cursor, and Prime Intellect—develops distributed systems that power thousands of concurrent simulations and AI copilots that accelerate environment creation. Backed by top-tier investors from OpenAI, Anthropic, xAI, and Thinking Machines, Proximal is pioneering the future of reinforcement learning for software engineering.

Company Size

1-10

Company Stage

Seed

Total Funding

N/A

Headquarters

San Francisco, California

Founded

2025

Your Connections

People at Proximal who can refer or advise you

Simplify Jobs

Simplify's Take

What believers are saying

  • Proximal is hitting the market exactly as reinforcement learning environments take off.
  • Training on real codebases (not static data) drives faster model improvement and clear user value.

What critics are saying

  • Long-horizon coding tasks are unsolved, giving room to set new methods and benchmarks.
  • Rising competition encourages deeper focus on hard engineering sims and strong moats.

What makes Proximal unique

  • Purpose-built for software-engineering reinforcement learning, not a retrofit of generic agent sims.
  • The team has previously built. coding agents and RL infra at OpenAI, Cursor, and Prime Intellect.
  • Backed by investors from OpenAI, Anthropic, xAI, and Thinking Machines.

Help us improve and share your feedback! Did you find this helpful?