Amazon SoC Performance Modeling Engineer Machine Learning Accelerators 

United States, California, Cupertino 

889499081

Custom SoCs (System on Chip) live at the heart of AWS Machine Learning servers, and our team builds C++ models for these ML accelerator chips ahead of silicon availability. The team currently develops functional models, but we’re expanding to include performance models for key components of the SoC. We’re looking for a modeling engineer to help us trail-blaze new technologies and architectures, while ensuring high quality with correlation against the design.As part of the ML accelerator modeling team, you will:
- Develop functional and/or performance models end-to-end, including model architecture, integration with other model or infrastructure components, testing, correlation, and debug
- Develop software which can be maintained, improved upon, documented, tested, and reused
- Drive model and modeling infrastructure performance improvements
Annapurna Labs, our organization within AWS, designs and deploys some of the largest custom silicon in the world, with many subsystems that must all be modeled, tested, and correlated with high quality. The model is a critical piece of software used in our SoC and SW stack development process. You’ll collaborate with many internal customers who depend on your models to be effective themselves, and you'll work closely with these teams to push the boundaries of modeling usage.You will thrive in this role if you:
- Are familiar with performance modeling of SoCs, ASICs, GPUs, or CPUs
- Are comfortable modeling in C++ and familiar with Python- Want to jump into an ML role, or get deeper into the details of ML at the system-level
Although we are building machine learning chips, no machine learning background is needed for this role. This role spans modeling of the ML and management regions of our chips, and you’ll dip your toes into both. You’ll be able to ramp up on ML as part of this role, and any ML knowledge that’s required can be learned on-the-job.This is a fast-paced role where you'll work with thought-leaders in multiple technology areas. You'll have high standards for yourself and everyone you work with, and you'll be constantly looking for ways to improve your software, as well as our products' overall performance, quality, and cost.

- 3+ years of non-internship professional experience writing functional or performance models
- Experience programming with C++
- Familiarity with SoC, CPU, GPU, and/or ASIC architecture and micro-architecture

- 3+ years of full software development life cycle, including coding standards, code reviews, source control management, build processes, and testing
- Experience developing and calibrating performance models for custom silicon chips
- Experience with writing benchmarks and analyzing performance
- Experience with PyTest and GoogleTest
- Familiarity with modern C++ (11, 14, etc.)
- Experience in multi-threaded programming, vector extensions, HPC, and QEMU
- Experience with machine learning accelerator hardware and/or software