Frontier Pretraining Data Engineer

Aimlroles; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

SpaceX is seeking Software Engineers to build data systems for frontier model training, focusing on data pipelines, crawling, and dataset preparation. You’ll drive end-to-end data quality, scalability, and reliability for training tokens and mixes used in model development.

The role spans Data Quality, Data Platform, and Crawling teams, with ownership of pipelines, experimentation, and collaboration with researchers and engineers to push model performance.

Qualifications

• Experience with large-scale data pipelines.
• Ability to work with researchers to improve data quality.
• Strong debugging and performance optimization skills.

Responsibilities

• Build high-throughput data pipelines with end-to-end traceability.
• Train and ship models that classify and filter data at extreme throughput.
• Design scaling experiments to validate data quality and dataset usefulness.
• Partner with Data Acquisition to source high-quality data and with training teams to improve loss and evals.
• Treat data quality as both a systems and a research problem.

Skills

Distributed systems

Data platforms

Web crawling

End-to-end ownership

Job description

SpaceX is seeking Software Engineers to build data systems for frontier model training, focusing on data pipelines, crawling, and dataset preparation. You’ll drive end-to-end data quality, scalability, and reliability for training tokens and mixes used in model development.

The role spans Data Quality, Data Platform, and Crawling teams, with ownership of pipelines, experimentation, and collaboration with researchers and engineers to push model performance.