Principal Scientist - Data Pipeline Engineer

What you'll do

• Own the end-to-end economics of turning billions of raw assets into training-ready data.
• Optimize pipeline DAGs to reduce costs by running cheap filters before expensive model stages.
• Improve throughput for VLMs, LLMs, and other models using batching, parallelism, and quantization.
• Distill large models into smaller, cheaper versions that preserve labeling and filtering quality.
• Build evaluation frameworks to ensure efficiency gains do not degrade training data quality.
• Encode data value into filtering, scoring, and selection systems for LLMs and diffusion models.
• Partner with data platform teams to run ML workloads cost-efficiently on shared infrastructure.

What we're looking for

• 10+ years of experience in ML systems, ML infrastructure, or applied ML engineering at a billions-of-assets scale.
• Proven track record of reducing costs for large-scale ML data pipelines across compute, storage, and data access.
• Deep expertise in optimizing inference for VLMs, LLMs, or other large models using batching, quantization, serving, and distillation.
• Experience with data curation and data quality for training generative or multimodal models.
• Proficiency in Python and a systems-level language such as C++, Rust, Go, or Java.
• Experience with distributed frameworks such as Ray or Spark.
• Bachelor's, Master's, or Ph.D. in CS, Engineering, ML, or a related field.
• Clear communication skills across modeling, infrastructure, and data teams.