Staff ETL Engineer

Unity Technologies•Washington, DC

About The Position

We are seeking a senior ML engineer to design and evolve the large-scale offline platform. This role focuses on building reliable infrastructure for generating training datasets, orchestrating ML workflows, and enabling efficient, distributed model training at scale. You will work closely with ML engineers and platform teams to ensure our pipelines can efficiently handle growing data volumes and increasingly complex training workloads. You will play a key role in shaping how model datasets are prepared as well as model training, validated, and delivered to distributed training systems, while ensuring the reliability, scalability, and performance of our offline ML platform.

Requirements

Strong experience building large-scale ML pipelines
Experience working with distributed computing frameworks such as Ray, Spark, Flink and familiarity in the Ray ecosystem (Ray Data, Ray Train) for distributed data processing and model training
Experience building infrastructure for training data generation, dataset preparation, or ML feature pipelines
Deep experience designing and operating production-grade data pipelines
Strong programming skills in Python and experience working with large-scale distributed workloads
Experience with modern data infrastructure (data lakes, warehouses, orchestration systems, streaming platforms)
Strong systems thinking, with the ability to reason about performance, scalability, reliability, and cost tradeoffs in distributed systems
Proven ability to lead technical direction and influence architectural decisions across teams without formal authority

Responsibilities

Design and operate large-scale data pipelines that generate training datasets used for machine learning training and experimentation
Develop infrastructure that supports distributed training workflows using technologies such as Pytorch, Ray Data, and Ray Train, etc.
Integrate ML pipelines with workflow orchestration systems (e.g., Flyte, Airflow, or similar) to enable reliable multi-stage training workflows
Improve reproducibility and observability of ML pipelines through dataset validation, monitoring, and automated testing
Optimize performance and resource utilization across distributed compute systems used for data processing and model training
Partner closely with ML engineers to enable efficient large-scale experimentation and model iteration
Lead architectural improvements to ensure our offline ML pipelines remain scalable, reliable, and cost-efficient