Junior Software Engineer

Responsibilities

• Design, build, and refine reinforcement learning (RL) environments used by frontier AI labs to train and evaluate coding models
• Own the full lifecycle of RL tasks, including ideation, implementing grading infrastructure, and running frontier models against tasks
• Analyze model failures to understand why and where they occur, iterating on tasks to ensure they are rigorous and fair
• Use coding agents heavily to execute tasks, directing them effectively and evaluating their output for subtle failures

Requirements

• Proficiency in Python
• Strong technical fundamentals
• Ability to work independently in a high-ownership environment
• Recent graduate or upcoming graduate

Preferred Qualifications

• Intuition for AI model behavior and the ability to anticipate where models might take shortcuts
• Significant experience working with frontier coding agents
• Motivation for solving problems that require both technical skill and creative cleverness

Benefits

• Base salary of $300,000
• Equity and performance bonuses
• Health, dental, vision, and life insurance

About the Company

Mechanize is a ~20 person team based in San Francisco that builds reinforcement learning environments for frontier AI labs. We focus on creating challenges that expose the gaps in current AI coding capabilities to help models improve. Mechanize is backed by prominent industry leaders including Patrick Collison, Nat Friedman, and Daniel Gross.