Machine Learning Engineer, Sequence Models (project Sequoia)

Amazon; New York (NY); On-site; USD 140,000 - 210,000; Full time

Job summary

Amazon's Sequence Models team is building scalable ML infrastructure to accelerate research-to-production for sequence models across Amazon's products. You will design and implement systems that support data pipelines, training, and model serving at petabyte scale.

As a Machine Learning Engineer, you will partner with scientists to move experiments into production, optimize GPUs, reduce latency, and establish automated data quality checks, monitoring, and on-call processes for reliable ML

Responsibilities

• Build and scale ML infrastructure across data processing, distributed training, and model serving. Optimize GPU utilization, training throughput, serving latency, and Infra costs.
• Own the data pipelines that feed model training, including ingestion of structured and unstructured inputs, schema evolution, backfills, and data quality checks across upstream sources.
• Partner with Applied Scientists to shorten the time from experiment to production.
• Evolve model serving and feature delivery to support continuous experimentation.
• Establish automated, repeatable processes for large-scale data analysis, model training, validation, and deployment.
• Own operational excellence for high-volume, low-latency production systems, including monitoring, alarming, troubleshooting, and on‑call.

Job description

Description

The Sequence Models team serves a centralized role developing sequence models to fundamentally understand Amazon's customers' journeys across all Amazon products (including Stores, Prime Video, Audible, Music, Twitch, etc.). These models will unlock new, actionable insights to optimize the customer experience, including when and how we serve ads. We leverage a host of scientific technologies to accomplish this mission, including Generative AI, classical ML, Causal Inference, Natural Language Processing, and Computer Vision.

As the Machine Learning Engineer on the team, you will deliver on our engineering vision to streamline the model development lifecycle from research to production, building ML infrastructure and establishing MLOps practices that enables rapid experimentation and deployment of ML models. You will invent and design new solutions to solve complex challenges that come with petabyte scale storage.

Key job responsibilities

• Build and scale ML infrastructure across data processing, distributed training, and model serving. Optimize GPU utilization, training throughput, serving latency, and Infra costs.
• Own the data pipelines that feed model training, including ingestion of structured and unstructured inputs, schema evolution, backfills, and data quality checks across upstream sources.
• Partner with Applied Scientists to shorten the time from experiment to production.
• Evolve model serving and feature delivery to support continuous experimentation.
• Establish automated, repeatable processes for large-scale data analysis, model training, validation, and deployment.
• Own operational excellence for high-volume, low-latency production systems, including monitoring, alarming, troubleshooting, and on‑call.