About the Role

This is a top-priority hire for a fast-growing, early-stage AI infrastructure company building the platform for reinforcement learning (RL) environments and post-training data. You'll own the automation of quality control (QC) for training data created by companies using the platform's infrastructure — a function that is central to the company's ability to scale.

You'll join a ~15-person engineering team made up of Olympiad medalists, AI startup founders, and published researchers. This role is ideal for someone who thrives in unstructured problem spaces, operates with strong autonomy, and has deep conviction about what makes training data genuinely good.

Visa sponsorship is available. The role is on-site in San Francisco, CA for US-based candidates, or Singapore for Southeast Asia–based candidates. Fully remote independent contractor arrangements are also considered, particularly for candidates based in Europe.

What You'll Do

• Automate QC for training data produced by companies using the platform's RL infrastructure.

• Build QC systems grounded in genuine understanding and human judgment — not heavy reliance on LLMs.

• Define and enforce quality standards for post-training datasets.

• Design experiments and metrics to evaluate and grade agent outputs.

• Partner with data vendors to debug quality issues, diagnose agent failure modes, and improve their data generation processes.

• Translate QC learnings into auditing systems, including sampling strategies and rule-based or model-assisted validation pipelines.

• Continuously integrate QC insights into infrastructure tooling and vendor-facing portals to reduce anomalies, inconsistencies, and edge cases.

What We're Looking For

Required (Dealbreakers):

• 2–4 years of experience in a research engineering or similar role with a focus on QC automation.

• Proficiency in Python, Docker, and Linux environments.

• Proven experience building scalable data validation pipelines and automated QA/QC systems end-to-end — without a fully prescribed roadmap.

Strongly Required:

• Experience working on benchmarks and evaluations for RL training data, including defining realistic tasks, reliable rubrics, and useful agent trajectories.

• Experience defining and measuring quality standards for training data.

• Ability to build QC systems based on human judgment and first-principles understanding rather than LLM-dependent workflows.

• Experience designing experiments and metrics to grade agent outputs.

• Experience collaborating with data vendors to debug quality issues, identify failure modes, and deliver actionable feedback.

Nice to Have:

• Knowledge of statistics and comfort designing QA/QC processes and measurement frameworks.

• Experience with existing benchmarks and constructing tasks for novel evaluations.

• Strong written and verbal communication skills for async collaboration across time zones.

• Genuine intellectual curiosity and a habit of asking clarifying questions that drive deeper understanding.

Compensation & Benefits

• Salary: $150,000 – $250,000 USD annually

• Equity participation in an early-stage, high-growth AI infrastructure company

• Visa sponsorship available for qualifying candidates

Location

• San Francisco, CA, USA — on-site (preferred for US-based candidates)

• Singapore — on-site (for Southeast Asia–based candidates)

• Fully remote / independent contractor — considered for candidates based in Europe or other locations outside the US and Southeast Asia