Software Engineer (SWE)
AI Evaluation Project Ownership Take end-to-end ownership of projects that push the frontier of AI evaluation. One week you might prototype a reinforcement learning environment from a research paper, the next deploy distributed experiments on Kubernetes, then improve a Next.js dashboard or build a Kafka pipeline for annotator analytics.
Responsibilities:
Design and build scalable systems: RL environments, APIs and human-in-the-loop platforms.
Collaborate with research, product and design to ship features quickly.
Write clean, maintainable code and contribute to documentation, code reviews and design discussions.
Solve real scalability and reliability challenges in the core infrastructure that powers data and evaluation for leading AI labs.
Work directly with AI lab researchers.
Requirements:
Two to five years as a full-stack engineer, with Next.js and React, Node.js or Python, and Kubernetes.
Experience building something end to end with a bias toward ownership.
Time at a high-growth, top-tier venture-backed startup, a reputable trading firm, or as a founding engineer.
A degree from a top-30 school in the US, Canada or Europe.
Comfort being client-facing with AI lab researchers.
Bonus Points:
Kafka, Redis and Elasticsearch.
Experience at a human data or data labeling AI company.
Location and Work ModelFull-time and on-site five days a week in the Financial District of San Francisco, with flexible hours.
Compensation$260,000-$290,000 base salary, plus uncapped commission of 2% of gross profit on projects you personally deliver and a 3% delivery-team pool paid quarterly, plus competitive equity. Target total cash compensation is $460,000-$690,000. Visa transfers are considered. Take-homes are paid.