ML Engineer Intern - Scalable Data Pipelines

Unity; Washington; On-site; USD 104,000 - 152,000; Part time

Benefits offered by this job

Health insurance

Stock ownership

Retirement/pension plans

Vacation and personal days

Wellbeing programs

Commuting subsidy

Job summary

Unity Vector builds an offline ML platform powering training data generation, ML workflows, and distributed model training at scale. You will help implement data pipelines and infrastructure for batch and streaming ML tasks, collaborating with researchers and engineers.

You’ll translate research ideas into production-grade systems, ensuring reliability and efficiency in large-scale ML pipelines and datasets.

Qualifications

• Bachelor’s degree in Computer Science, Machine Learning, Systems, or a related field

Responsibilities

• Build and maintain data pipelines that generate training datasets for ML models and experimentation
• Contribute to infrastructure that supports distributed training workflows (e.g., PyTorch, Ray)
• Work with workflow orchestration tools (Airflow, Flyte, or similar) to support multi-stage ML pipelines
• Improve reproducibility and reliability through dataset validation, monitoring, and testing
• Partner with ML engineers to support experimentation and model iteration
• Help optimize performance and efficiency across data processing and training systems
• Contribute to the evolution of our offline ML platform architecture as it scales

Skills

Python

PyTorch

TensorFlow

Distributed systems

Data pipelines

Spark

Ray

Education

Bachelor's degree in CS/ML/Systems

Tools

Airflow

Flyte

Spark

Ray

Job description

Unity Vector builds an offline ML platform powering training data generation, ML workflows, and distributed model training at scale. You will help implement data pipelines and infrastructure for batch and streaming ML tasks, collaborating with researchers and engineers.

You’ll translate research ideas into production-grade systems, ensuring reliability and efficiency in large-scale ML pipelines and datasets.