Join an engineering team at an early-stage AI infrastructure company focused on reinforcement learning environments and evaluation. You will build quality-control systems for training data, helping ensure that data and evaluations are reliable as demand grows.
Automate quality control for training data produced using the company's infrastructure.
Define and enforce clear quality standards using human judgment and careful analysis, rather than relying heavily on large language models.
Design experiments and metrics to assess agent outputs and evaluation quality.
Work with data vendors to investigate quality issues, diagnose agent failure modes, and improve data-generation processes.
Turn quality-control findings into auditing systems, sampling strategies, and rule-based or model-assisted validation pipelines.
Improve infrastructure tools and vendor-facing workflows to reduce anomalies, inconsistencies, and edge cases.
Two to four years of experience in a research engineering or similar role focused on quality-control automation, plus a degree in a technical field.
Proficiency with Python, Docker, and Linux.
Experience independently building scalable data-validation pipelines and end-to-end automated QA or QC systems.
Background in benchmarks and evaluations for reinforcement learning data, including realistic task design, reliable rubrics, and useful training trajectories.
Experience defining and measuring training-data quality, and designing experiments, metrics, and quality processes.
Strong statistical knowledge and the ability to communicate clearly across time zones and work independently in an unstructured startup environment.
Base salary: $100,000 to $170,000 USD annually. Visa sponsorship is available.
On-site in Singapore, Singapore.