Software Development Manager, Amazon Robotics
Posted Updated
Description
The Amazon Robotics Pack team builds the software and robotic systems that pack nearly every customer order Amazon ships. Our work is on the critical path: when a pack machine or station goes down, shipments stop. We run thousands of machines, stations and robotic workcells across hundreds of fulfillment centers on several continents. Each one has to perform reliably every shift, for both the associates who work alongside it and the customers waiting on their orders.
To do that at scale, we're building physical AI that runs at the edge: software that sees, reasons and acts on the machine itself, in real time. Computer vision and machine learning on edge compute verify items, select the right box and catch defects before a package leaves the station. The same systems watch the machines themselves. They track throughput, jam rates, sensor and actuator signals, and vision-model accuracy, so we know which machines are healthy, which are degrading and why.
When something goes wrong, our systems should spot it first, diagnose it and guide the fix. That means turning machine telemetry into clear, step-by-step guidance for associates and Reliability Maintenance Engineering (RME) technicians, so they can restore a machine in minutes instead of hours. Every failure feeds back into our models and designs, so a fault fixed once at one site helps prevent it at every site. The hard part is doing this across a large, mixed fleet, with low latency on the floor and consistent quality everywhere.
A day in the life
1) Lead, hire and grow a team of software engineers
2) Own the roadmap for software that runs pack machines, stations and robotic workcells
3) Integrate computer vision, sensors and robotics into reliable real-time systems
4) Turn every failure into fleet-wide improvements to models, designs and runbooks
5) Solve optimization problems in packaging selection, workload balancing and throughput
6) Drive operational excellence: safe deployments, fast mitigation, root-cause COEs
7) Launch new robotics and automation into live fulfillment sites
8) Work backwards from associates, RME, site leaders and end customers
Influence partner teams across hardware, science, operations and product
About the team
Our team knows fulfillment end to end. We understand how an order moves through Amazon's network around the world, from inbound to the moment a package leaves the building. That broad view lets us make good decisions quickly and see how a change at one pack station affects the rest of the network.
We have a strong bias for action, backed by mechanisms that keep us focused on delivering. We don't let obstacles stop us. When we hit a blocker, we treat it as something to learn from, work through it and keep moving. That's the attitude we bring every day: we get results, whatever the challenge.
To do that at scale, we're building physical AI that runs at the edge: software that sees, reasons and acts on the machine itself, in real time. Computer vision and machine learning on edge compute verify items, select the right box and catch defects before a package leaves the station. The same systems watch the machines themselves. They track throughput, jam rates, sensor and actuator signals, and vision-model accuracy, so we know which machines are healthy, which are degrading and why.
When something goes wrong, our systems should spot it first, diagnose it and guide the fix. That means turning machine telemetry into clear, step-by-step guidance for associates and Reliability Maintenance Engineering (RME) technicians, so they can restore a machine in minutes instead of hours. Every failure feeds back into our models and designs, so a fault fixed once at one site helps prevent it at every site. The hard part is doing this across a large, mixed fleet, with low latency on the floor and consistent quality everywhere.
A day in the life
1) Lead, hire and grow a team of software engineers
2) Own the roadmap for software that runs pack machines, stations and robotic workcells
3) Integrate computer vision, sensors and robotics into reliable real-time systems
4) Turn every failure into fleet-wide improvements to models, designs and runbooks
5) Solve optimization problems in packaging selection, workload balancing and throughput
6) Drive operational excellence: safe deployments, fast mitigation, root-cause COEs
7) Launch new robotics and automation into live fulfillment sites
8) Work backwards from associates, RME, site leaders and end customers
Influence partner teams across hardware, science, operations and product
About the team
Our team knows fulfillment end to end. We understand how an order moves through Amazon's network around the world, from inbound to the moment a package leaves the building. That broad view lets us make good decisions quickly and see how a change at one pack station affects the rest of the network.
We have a strong bias for action, backed by mechanisms that keep us focused on delivering. We don't let obstacles stop us. When we hit a blocker, we treat it as something to learn from, work through it and keep moving. That's the attitude we bring every day: we get results, whatever the challenge.