Description

Build the infrastructure that builds Amazon. Join the Build Compute Team and work on distributed systems processing 25 million tasks daily across 12K hosts - powering every software deployment from Alexa to AWS.

As an SDE II, you'll solve challenging distributed systems problems at massive scale - from optimistic concurrency and circuit breakers to AI-powered optimization. We're building next-generation infrastructure on NAWS architecture to harden security boundaries, scaling to 120K+ build instances across Amazon.

What You'll Do:

- Own critical components of Amazon's Tier-1 build infrastructure

- Design and implement core components to power large scale NAWS-based build fleet

- Leverage AI/ML for intelligent task scheduling and resource optimization

- Scale systems handling 2-3x growth while maintaining high performance and availability

- Debug complex distributed systems issues

- Mentor teammates on operational excellence and system architecture

We're looking for engineers who thrive on making systems faster, more reliable, and more secure at Amazon scale.


Key job responsibilities
- Design, implement, and maintain distributed systems for Brazil Build System (Tier-1) processing 25 million tasks daily

- Build next-generation NAWS-based infrastructure to harden security boundaries, scaling to 120K+ build instances

- Operate and continuously improve task dispatching (NBTM) and build execution (DockerBuildFleetNode) with 99.99% availability

- Participate in on-call rotation, solving difficult customer and operational problems while addressing root causes over short-term workarounds

- Deliver high-quality, maintainable code with comprehensive test automation (unit and integration testing)

- Review code, mentor teammates, and train new engineers on system architecture, testing practices, and operational procedures

- Work with legacy and modern technologies, making appropriate trade-offs to leave systems better and easier to maintain

- Communicate status and changes to stakeholders, keeping all informed of updates to plans and deliverables

About the team
We power Amazon's Brazil Build System - a Tier-1 platform processing 25 million tasks daily across 12K hosts (384K cores). Every AWS and SDO engineer depends on us to transform source code into deployment-ready artifacts.

Our stack: NeverBuildTwiceManager (NBTM) handles distributed task orchestration with JournalDB transactions, optimistic concurrency. DockerBuildFleetNode manages containerized build execution with sophisticated resource allocation and real-time capacity management across the fleet.

The challenge: maintain 99.99% availability at massive scale while continuously optimizing performance, mitigating risks, and solving complex distributed systems problems. Your code directly impacts thousands of developers and Amazon's deployment velocity.