3D Perception Lead for Robotics (Applied Science Manager)

Amazon; Seattle (WA); On-site; USD 184,000 - 249,000; Full time

Benefits offered by this job

Medical, Dental, and Vision Coverage

401(k) Plan

Parental Leave Options

Job summary

Amazon Robotics’ Vulcan Stow team seeks an Applied Science Manager to lead a world-class perception group. The role drives 3D scene understanding and real-time perception for robots, delivering robust, low-latency outputs in fulfillment centers and beyond.

You will guide researchers and engineers in transformer-based architectures, generative 3D models, and scalable data pipelines, collaborating across match & affordances, motion planning, hardware, and software teams to ship production-ready

Qualifications

• PhD or Master's degree with 4+ years of industry or academic research.
• 6+ years of applied research experience.
• 3+ years of management experience in scientists or ML engineers.
• Experience building ML models or algorithms for business applications.
• Experience with deep learning libraries (PyTorch, TensorFlow, MXNet).
• Knowledge of software/hardware development life cycle and practices.

Responsibilities

• People Leadership: recruit, motivate, and coach a diverse team of 3D vision and perception specialists.
• Technical Vision: set roadmaps for 3D scene understanding, depth estimation, and multi-view fusion.
• Cross-functional Collaboration: partner with motion planning, hardware, fulfillment, and ML infra teams.
• Delivery Excellence: manage timelines, resources, and deliverables; contribute to architecture decisions.
• Problem Solving: lead post-mortems to identify root causes of perception regressions and latency issues.

Skills

People management

3D computer vision

Transformer-based architectures

Generative modeling

Edge deployment

Education

PhD or Master’s with 4+ years research

Tools

PyTorch

TensorFlow

MxNet

Job description

Amazon Robotics’ Vulcan Stow team seeks an Applied Science Manager to lead a world-class perception group. The role drives 3D scene understanding and real-time perception for robots, delivering robust, low-latency outputs in fulfillment centers and beyond.

You will guide researchers and engineers in transformer-based architectures, generative 3D models, and scalable data pipelines, collaborating across match & affordances, motion planning, hardware, and software teams to ship production-ready