Member of Technical Staff - ML Research Engineer, Pre-Training Data

Highlights

• Own pre-training data mixtures and curricula for Liquid Foundation Models.
• Run proxy experiments and ablations to measure what data changes help.
• Build synthetic data systems to cover what web-scale data can't.

What they offer

• Competitive base salary plus equity
• Unlimited PTO plus company-wide Refill Days
• Comprehensive health coverage, details vary by location
• Paid leave for pregnancy, parenting, medical needs and family care

Ideal candidate

A researcher-engineer who owns ambiguous data problems end to end, leads with hypotheses, and cares how data shapes model behavior. You thrive on small teams turning experiments into shipped pre-training runs.

Requirements

• Deep hands-on pre-training data work: web-scale sourcing and curation, synthetic data generation, mixing and curriculum design, deduplication and decontamination
• Ability to reason from data decisions to training effects across model size, architecture and training dynamics
• Designed experiments to test data changes with baselines, scaling ladders, proxies and held-out loss
• Built and owned production data pipelines, not only research prototypes
• Solid ML fundamentals with experience training and evaluating models in Python (PyTorch preferred)
• 3+ years relevant experience with M.S., 1+ year with Ph.D., or 5+ years with B.S.