Senior ML Engineer, Generative AI Platform (EKS/GPU)

Prime Video & Amazon MGM Studios; New York (NY); On-site; USD 185,000 - 262,000; Full time

Job summary

Prime Video and Amazon MGM Studios seeks a Machine Learning Engineer for CreativeFlux, an ML platform powering Nara for AI-native content creation. You will build and operate training and inference infra on EKS, deploy multi-GPU workloads, and collaborate with Applied Scientists to scale new architectures.

You will optimize GPU utilization, integrate with DCC tools, and own model endpoints' health, monitoring, and on-call responses while advancing generative ML and distributed training

Qualifications

• 5+ years of professional software development experience.
• 5+ years of programming in at least one language.
• 5+ years leading design/architecture of systems.
• Experience as mentor, tech lead, or leading an engineering team.

Responsibilities

• Build and operate ML training and serving infrastructure on the CreativeFlux platform.
• Design and deploy ML workloads on EKS with GPU node groups and autoscaling.
• Integrate generative ML models into production inference pipelines with scientists.
• Build training/fine-tuning pipelines on EKS and SageMaker with data prep and checkpointing.
• Improve GPU utilization and latency via batching, quantization, and Kubernetes scaling.
• Develop automated evaluation pipelines and human-in-the-loop review.
• Integrate with DCC tools like Adobe Animate, After Effects, Maya, and Storyboard Pro.
• Own operational health of model endpoints, monitoring, alerts, and on-call rotations.
• Write clean, well-tested code and participate in reviews.
• Stay current with generative ML and distributed training techniques.

Skills

Non-internship software development

Programming experience

Design/architecture leadership

Mentor/tech lead experience

Education

Bachelor's degree in computer science or equivalent

Tools

Kubernetes

EKS

Job description

Prime Video and Amazon MGM Studios seeks a Machine Learning Engineer for CreativeFlux, an ML platform powering Nara for AI-native content creation. You will build and operate training and inference infra on EKS, deploy multi-GPU workloads, and collaborate with Applied Scientists to scale new architectures.

You will optimize GPU utilization, integrate with DCC tools, and own model endpoints' health, monitoring, and on-call responses while advancing generative ML and distributed training