Senior AI Training Infra Engineer — GPU Clusters

Designworks Talent; Bellevue (WA); Hybrid; USD 180,000 - 300,000; Full time

Benefits offered by this job

Medical, dental, vision insurance

401(k) with company match

Paid holidays

Job summary

Designworks Talent is seeking an AI Training Infrastructure Engineer for a hybrid role in the Bellevue, WA area. You will build and scale distributed training infrastructure powering large AI models across multi-node GPU clusters and production pipelines.

Work with platform, orchestration, performance, and ML teams to improve reliability, efficiency, and scalability. Responsibilities include fault-tolerance, checkpointing, recovery, and developing developer tools to streamline AI research

Qualifications

• Hands-on experience building distributed training systems.
• Experience with large AI models and production ML pipelines.
• Ability to optimize multi-node GPU training for reliability and efficiency.
• Proven track record delivering scalable ML infrastructure solutions.

Responsibilities

• Build and scale distributed training infrastructure supporting large AI models across GPU clusters.
• Design and improve systems for training reliability, efficiency, and resource utilization.
• Develop fault tolerance, checkpointing, and recovery for large-scale training.
• Integrate training systems with production ML pipelines in cross-functional teams.
• Diagnose and resolve bottlenecks affecting throughput, stability, and cost efficiency.
• Create tools and automation to improve researcher and engineer productivity.
• Establish best practices for training infra and platform reliability.
• Contribute to evolving the AI infra platform as an early engineering team member.

Skills

Distributed training

GPU clusters

Python

Multi-node systems

Production ML pipelines

Tools

PyTorch Distributed

DeepSpeed

Kubernetes

GPU virtualization

Job description

Designworks Talent is seeking an AI Training Infrastructure Engineer for a hybrid role in the Bellevue, WA area. You will build and scale distributed training infrastructure powering large AI models across multi-node GPU clusters and production pipelines.

Work with platform, orchestration, performance, and ML teams to improve reliability, efficiency, and scalability. Responsibilities include fault-tolerance, checkpointing, recovery, and developing developer tools to streamline AI research