Machine Learning Engineer II, AWS Just-Walk-Out Science Team

Amazon Web Services (AWS); Seattle (WA); On-site; USD 120,000 - 180,000; Full time

Job summary

An innovative firm is seeking a Machine Learning Engineer II to join their cutting-edge Just Walk Out Science Team. In this exciting role, you will leverage advanced machine learning techniques and Amazon's vast data resources to develop algorithms that enhance customer experiences. Collaborate with a team of experts to design and maintain large-scale distributed training systems, ensuring the reliability and performance of AI applications. This position offers a unique opportunity to work at the forefront of technology, contributing to a revolutionary retail concept that eliminates traditional checkout processes. If you're passionate about AI and eager to make a significant impact, this role is perfect for you.

Qualifications

• 3+ years of software development experience with coding standards and operations.
• Experience with system design, reliability, and scalability.

Responsibilities

• Integrate advanced model architectures into training pipelines.
• Design large-scale distributed training systems for multi-modal models.

Skills

Python

PyTorch

Deep Learning Frameworks

System Design

Machine Learning Operations

Education

Bachelor's in Computer Science

Master's or PhD in Computer Science

Tools

MMEngine

Megatron-LM

GPU Optimization Techniques

Job description

Machine Learning Engineer II, AWS Just-Walk-Out Science Team

Join to apply for the Machine Learning Engineer II, AWS Just-Walk-Out Science Team role at Amazon Web Services (AWS)

Job Description

As part of the AWS Solutions organization, we aim to provide innovative business applications leveraging Amazon’s expertise, used by millions worldwide to manage daily operations. Our goal is to accelerate customer businesses through intuitive and differentiated technological solutions that address enduring business challenges. We combine vision with curiosity and real-world experience to build turnkey solutions, becoming trusted partners for customers who prefer buying over building.

The Team

Just Walk Out (JWO) is a revolutionary retail concept—no lines, no checkout. Customers use the Amazon Go app to enter, pick their items, and leave; charges are processed automatically. Our technology, including computer vision, sensor fusion, deep learning, and foundation models, powers this seamless experience. We are committed to innovation and pushing the boundaries of computer vision, AI, and hardware design. Our team designs autonomous AI agents capable of complex scene understanding, object permanence reasoning, and customer behavior prediction.

Key Responsibilities

• Collaborate with Applied Scientists to integrate advanced model architectures into training pipelines and auto-labeling processes.
• Process large datasets, scale machine learning models, optimize GPU utilization, and improve training workflows (kernel fusion, mixed-precision, gradient accumulation, offloading, parallelization).
• Design and maintain large-scale distributed training systems for multi-modal foundation models supporting autonomous retailing.
• Develop monitoring and debugging tools to ensure training workflow reliability and performance.
• Work with engineers and scientists to explore new design approaches, prototype technology, and solve complex problems.

Daily Responsibilities

Lead development of algorithms and models for hardware-accelerated training, impacting customer products and services. Leverage Amazon’s data and computing resources to accelerate AI applications. Influence team strategy, system architecture, and best practices within an Agile environment.

Basic Qualifications

• 3+ years of professional software development experience with coding standards, code reviews, source control, build, testing, and operations.
• 2+ years of experience in system design or architecture, including reliability and scalability.
• Proficiency in Python or related language.
• Hands-on experience with PyTorch, deep learning frameworks like MMEngine or Megatron-LM, and large-scale deep learning operations.
• Knowledge of visual-language models, multi-modal AI, pre-training/post-training techniques, and performance analysis tools.

Preferred Qualifications

• Master's or PhD in computer science or related field.
• Over 1 year of experience deploying or optimizing ML models, with strong skills in scalable distributed GPU training frameworks and open-source LLM contributions.
• Experience with large-scale multimodal models, GPU optimization techniques, and multi-modal learning tasks.

We are an equal opportunity employer. For accommodations during the hiring process, visit the provided Amazon link. Compensation varies by location and experience, with full benefits and potential equity or sign-on bonuses. Apply via our careers site.