AWS runs one of the largest networks in the world. We design and implement its network devices and their operating system, NetOS. Every AWS service runs on this network, and every packet in the cloud touches our code! Nobody notices a healthy network. Operators need to know within seconds when a link degrades in one data center, and the first signal comes from software on the device. We write that software. Our team, NetOS Observability, builds the agent that collects metrics, events, and logs from every NetOS device and publishes them to the network telemetry platform. Fleet health systems, SDN controllers, and network operators consume what we ship. We are moving the fleet off a legacy daemon and onto a brand new Rust agent. In 2027, we take it to every supported platform, retire the legacy collectors, and let builders publish a new metric fleet-wide in two weeks. You will own that agent end to end, from its design to its release across the fleet to its behavior in production. The decisions that shape the next decade of network telemetry are open now. We need your help to craft the future. Key job responsibilities - Design and implement the NetOS observability agent in Rust on Linux (ARM64). - Own the telemetry pipeline from device to gateway, including transport, authentication, schema, backpressure, and fleet-scale correctness. - Migrate consumers off the legacy daemon with adapters that preserve their data shape, then retire the daemon. - Build the test and release machinery that ships the component to the fleet on its own cadence. Every release passes its test gate. - Instrument your own systems, define their SLOs, and carry the on-call rotation for them. - Write design documents and review other engineers' designs. Name the dependency early when another team's roadmap gates yours, and negotiate the date. A day in the life You write Rust and test on real routers in our lab. You build for the component teams who define their telemetry in a schema we own, and for the fleet health systems, controllers, and operators who consume it. Every two weeks you settle the device-to-cloud contract with the telemetry platform team. You add a metric to the schema, bring a legacy consumer onto the new pipeline, and run the result on a router under its memory budget. AI agents draft, review, and test alongside you. You ship the component every few weeks and carry its on-call. About the team Our team sits inside NetOS Managed Fleet, which delivers health, observability, and management services for the NetOS fleet. We are growing the team in Raleigh, NC. You work directly with senior engineers and leadership across AWS Networking, and you give technical guidance as the team grows. We own what we ship, and we carry the on-call for the agent in production. We value clear design writing, incremental delivery, and property-based tests.