Senior Vision System Engineer, Amazon Robotics

We're building the next generation of vision and perception systems that give machines the power to see, understand, and act, and we need a Vision Systems Engineer to lead the charge from customer problem to deployed solution. If you love working at the intersection of sensors, machine learning, and compute hardware, and you want your work to directly shape products that operate at massive scale, this is the role for  this position, you'll be the bridge between our customers and our engineering teams.

You'll gather real world requirements, determine the right combination of sensors and ML models, define what compute hardware is needed to run them, and hand off clear specs to hardware teams for execution. This is a high impact, cross functional role where your technical judgment drives what we build and how we build it.

You won't be siloed into one discipline. You'll operate across the full perception stack, from the optics on the sensor to the neural network running on an edge device to the customer conversation where it all starts. Your work directly shapes what gets built, and you'll see the impact end to end.

You're someone who is equally comfortable talking to customers about their challenges and diving into datasheets to find the right component. You can explain the difference between a depth ML model and a classical stereo algorithm to a non technical audience, and you can turn around and have a detailed conversation with a hardware engineer about power budgets and memory bandwidth.

You connect the dots between sensing, computation, and real world performance.

Key job responsibilities
- Engage directly with customers to understand perception and sensing requirements, and translate those into clear, quantifiable engineering specifications
- Evaluate and recommend sensor technologies (cameras, depth sensors, barcode readers) for each application, performing trade off analysis on resolution, frame rate, field of view, cost, and reliability
- Define the ML model requirements for perception tasks such as depth estimation, object detection, damage detection, and barcode reading
- Profile compute resource needs (processing power, memory, accelerators, power consumption) for running inference models and determine optimal deployment locations (edge, cloud, hybrid)- Communicate hardware requirements including power budgets, memory bandwidth, thermal constraints, and I/O interfaces to hardware engineering teams
- Author system architecture documents, interface control documents, and technical trade studies
- Drive cross functional alignment between software, hardware, and product teams to ensure customer requirements are met end to end
- Support lab and field activities related to sensor integration, system validation, and performance characterization
- Identify technical risks early in the design cycle and propose mitigation strategies that balance performance, cost, and scheduleA day in the life

Your day starts with a customer call. They need to add damage detection to an inspection station but don't know what sensor or compute platform will work. You ask the right questions, dig into the constraints, and start shaping a technical approach.

Later, you're reviewing inference benchmarks to recommend whether the current edge board can handle another ML model, or if hardware needs to spec something more robust. By afternoon, you're presenting a sensor trade study to the hardware and software teams, walking through the pros and cons of two candidates.

No two days look the same. Some days you're drafting architecture documents. Others you're visiting a customer site to see challenges firsthand. You're the connective tissue between what customers need and what engineering builds.

Amazon offers a full…