Perception Science Manager, 3D Vision for Robotics

Amazon Inc.; Seattle, Northern (WA, KY); Hybrid; USD 184,000 - 249,000; Full time

Benefits offered by this job

Medical Coverage

Dental Coverage

Vision Coverage

401(k) Plan

PTO

Job summary

Amazon Robotics is seeking an Applied Science Manager to lead the Perception team. You will drive technical vision for 3D scene understanding, coordinate with cross-functional teams, and deliver perception models for real robots in fulfillment centers.

You will mentor scientists and engineers, guide model development from research to production, and optimize inference latency while maintaining robust accuracy in dynamic, real-world environments.

Qualifications

• PhD or Master’s degree required (or equivalent).
• 4+ years of research experience in industry or academia.
• 6+ years of applied research experience.
• 3+ years of scientists or ML engineers management experience.
• Experience building ML models or algorithms for business applications.
• Experience with deep learning libraries (PyTorch, TensorFlow, MXNet).
• Knowledge of software/hardware/networks development lifecycle and CI/CD practices.

Responsibilities

• Lead a team of applied scientists and engineers focused on 3D perception for robotics.
• Set technical vision and roadmaps for 3D scene understanding and perception models.
• Collaborate with cross-functional teams across match & affordances, motion planning, and fulfillment operations.
• Oversee training pipelines, edge deployment optimization, and model evaluation.
• Manage project timelines, resources, and deliverables.

Skills

People management

3D computer vision

Transformer-based architectures

Generative modeling

Edge deployment

Education

PhD or Master’s degree

Tools

PyTorch

TensorFlow

MxNet

Job description

Amazon Robotics is seeking an Applied Science Manager to lead the Perception team. You will drive technical vision for 3D scene understanding, coordinate with cross-functional teams, and deliver perception models for real robots in fulfillment centers.

You will mentor scientists and engineers, guide model development from research to production, and optimize inference latency while maintaining robust accuracy in dynamic, real-world environments.