Senior ML Engineer, 3D Data & Synthetic Pipelines

SportsTechJobs; San Mateo (CA); On-site; USD 243,000 - 295,000; Full time

Benefits offered by this job

Equity compensation

Job summary

Roblox seeks a Senior/Principal Machine Learning Engineer in San Mateo, CA to architect and scale data pipelines for 3D and video foundation models. You will own ingestion, cleaning, and synthetic data generation for multi-modal datasets at petabyte scale, collaborating with researchers to push models from experiments to production.

The role emphasizes high-throughput inference, robust evaluation, and efficient resource use, with equity compensation and a path to leadership within the Foundation

Qualifications

• 8+ years of experience as a data systems engineer, preferably with 3D and video foundation models.
• Experience building scalable ML data pipelines for batch and real-time environments.
• Proficient in Python; comfortable with C++ for performance-critical components.
• Experience with cloud data platforms and distributed processing (e.g., Spark, Ray, Kubeflow, S3).

Responsibilities

• High-Scale Data Orchestration for multi-modal datasets spanning petabytes.
• Synthetic data generation to scale multi-modal datasets.
• Research-to-Production bridge for synthetic data creation pipelines.
• Scalable evaluation frameworks with automated metrics and human-in-the-loop interfaces.
• Model deployment and API architecture for high-throughput inference.
• Autonomous SOTA tracking through literature reviews and experiments.
• Resource efficiency and observability to maximize GPU utilization.

Skills

Python proficiency

Distributed data pipelines

Large-scale data handling

ML data pipelines

Team leadership

C++ knowledge

3D/Video data

Cloud data platforms

Education

Bachelor's degree or equivalent in CS/CE

Tools

Weights & Biases

MLflow

Spark

Ray

Kubeflow

S3

Job description

Roblox seeks a Senior/Principal Machine Learning Engineer in San Mateo, CA to architect and scale data pipelines for 3D and video foundation models. You will own ingestion, cleaning, and synthetic data generation for multi-modal datasets at petabyte scale, collaborating with researchers to push models from experiments to production.

The role emphasizes high-throughput inference, robust evaluation, and efficient resource use, with equity compensation and a path to leadership within the Foundation