Sr. Firmware Engineer, Annapurna Labs, Machine Learning Acceleration - Power and Performance

Amazon•Cupertino, CA

50d

About The Position

AWS Utility Computing (UC) provides product Annapurna Labs (our organization within AWS UC) designs silicon and software that accelerates innovation. Customers choose us to create cloud solutions that solve challenges that were unimaginable a short time ago—even yesterday. Our custom chips, accelerators, and software stacks enable us to take on technical challenges that have never been seen before, and deliver results that help our customers change the world. We are seeking a Senior Firmware Engineer to join our Power Architecture team, developing firmware algorithms for power and performance management on ML Acceleration Chips. In this role, you will design and implement intelligent control algorithms, optimization strategies, and real-time decision-making systems that maximize performance while managing power and thermal constraints. You will develop sophisticated firmware that monitors system state, makes dynamic trade-offs between power and performance, and implements adaptive control policies. To enable this work, you will also build instrumentation and tracing capabilities that provide the telemetry needed to develop, tune, and validate your algorithms, with collected data optionally post-processed using cloud-based analytics. You will work closely with power architects and hardware teams to understand silicon capabilities, implement low-level control mechanisms, and create the algorithms and tooling that deliver optimal system behavior. This position is ideal for firmware engineers who enjoy solving algorithmic challenges in resource-constrained environments, working close to hardware, and building systems that intelligently manage complex trade-offs in real-time.

Requirements

5+ years of non-internship professional software development experience
Experience as a mentor, tech lead or leading an engineering team
Bachelor's degree in computer science, electrical engineering, or related field
Strong firmware or embedded systems development experience
Proficiency in C/C++ for systems programming with strong foundation in algorithms and data structures
Experience implementing efficient algorithms in resource-constrained, real-time environments
Experience with hardware interfaces, instrumentation, or performance monitoring
Strong debugging skills with hardware-software systems
Experience building developer tools or instrumentation frameworks

Nice To Haves

Experience developing control algorithms, optimization algorithms, or state machines in firmware
Experience with power management algorithms, thermal control policies, or dynamic performance optimization
Background in tracing frameworks, telemetry systems, or performance analysis
Understanding of algorithmic complexity and optimization techniques for embedded systems
Familiarity with hardware performance counters, on-chip monitoring, or hardware debug interfaces
Experience with data collection pipelines and scripting (Python, shell) for algorithm validation
Understanding of ML training/inference workloads and their performance characteristics
Takes strong ownership, works effectively in ambiguous situations, demonstrates a bias for action while consistently delivering impactful results

Responsibilities

Design and implement firmware algorithms for power management, thermal control, and performance optimization on ML acceleration hardware
Develop real-time control policies and state machines that dynamically balance power, thermal, and performance constraints
Create optimization algorithms for resource allocation, frequency/voltage scaling, and workload scheduling
Implement efficient data structures and algorithms suitable for embedded, resource-constrained environments
Design and implement on-device tracing and telemetry collection systems to support algorithm development and validation
Build developer tools and data pipelines for metric collection, analysis, and visualization of algorithm behavior
Implement low-overhead instrumentation that minimizes impact on workload performance
Collaborate with hardware architects to understand hardware capabilities and identify optimal instrumentation points
Develop automated testing and validation workflows; integrate with optional cloud-based analytics pipelines
Own firmware code quality through rigorous testing, debugging, and validation on hardware