Senior Amazon Software Engineer Jobs in Boston, MA

Amazon Robotics is seeking a Senior Embedded Software Engineer with deep Rust expertise to help lead development of safety-rated sensor processing firmware for one of the world's largest autonomous robot fleets. You will set technical direction for major subsystems of the Safety Perception stack (real-time safety applications, sensor processing pipelines, and system interfaces on heterogeneous multi-core MCU/CPU/GPU platforms) that let robots in Amazon facilities worldwide detect and respond to obstacles safely and reliably. You will own the team's hardest end-to-end problems, framing solutions where the business need is clear but the technical strategy is not, and raise the bar for the engineers around you.
We are looking for a technical leader who is deeply proficient in Rust (or strong in C/C++ and eager to work in Rust), is fluent across the boundary between RTOS and Linux on the same chip, has substantial experience with sensor data acquisition or signal processing, and can architect systems that meet the rigor of IEC 61508 functional safety certification

You will influence beyond your immediate team, shaping designs, resolving escalations, mitigating long-term one-way-door risks, and mentoring engineers, all while collaborating closely with hardware, functional safety, and navigation teams to deliver firmware where correctness is not optional.
Key job responsibilities
The ideal candidate is a force multiplier who sets technical direction across teams and raises the bar for associate safety. They will:
- Lead the design and architecture of major safety-rated and non-safety-rated firmware subsystems in Rust on heterogeneous multi-core platforms, where real-time safety code and Linux services coexist on the same chip.
- Set technical direction across the full sensor processing pipeline (driver integration, frame acquisition, point cloud generation, collision detection) and influence peer hardware, safety, and navigation teams.
- Own the team's hardest end-to-end problems, framing solutions where the business problem is known but the technical strategy is not.
- Drive perception features that let robots detect and respond to obstacles under IEC 61508 certification, ensuring data integrity through black-channel diagnostics.
- Raise the technical bar through design and code reviews, and mentor engineers so the team is more reliable because of your presence but does not require it.
- Produce direction-setting design documents and safety analysis artifacts that meet the rigor of third-party certification and witness testing.
- Lead debugging and root-cause analysis end-to-end, from raw sensor data to safety responses, on real robots at test facilities and in production.
- Leverage AI coding assistants as a force multiplier and lead the team to do the same.
A day in the life
No two days are the same. You might spend the morning aligning peer teams around a subsystem architecture in a design review, then head to the test floor after lunch to lead a cross-functional debug with electrical and QA engineers, determining whether an issue is firmware, hardware, or a faulty sensor

That afternoon you profile a pipeline to close a latency budget, then pair with another engineer to unblock their design. Our engineers take complete ownership (from interrupt handlers to CI/CD pipelines to on-site data collection) and frequently enter new areas to meet our customers' needs. As a senior engineer, you help chart where the team goes next.
About the team
The Safety Perception team is a group of passionate embedded software engineers who apply advances in robotics and sensor processing to real-world safety challenges

We write embedded Rust on platforms where correctness is not optional, and we own the full stack: from low-level driver integration to system APIs to the test infrastructure that proves our firmware works. We take our responsibility for safety seriously while challenging ourselves to ship the right product at the performance we need. We pay attention to the details so others do not have to, making systems intuitive enough that safety does not need a second thought.