Sr. Manager - SOC Virtual Platform Modeling, ML Acceleration - Annapurna Labs

Custom ML acceleration SoCs (System on Chips) are the brains behind AWS’s Trainium Machine Learning servers. Our team builds functional models of these SoCs and the servers they power. Our models are used by AWS internal teams for silicon verification and to left‑shift software development (as a virtual platform). We are looking for a leader who can quickly scale the team and begin executing on a performance model‑someone who can drive both the team vision and hiring.

Great candidates will have deep hands‑on modeling experience, with a startup‑like can‑do mentality - the team is nimble, and we need a leader who can jump in and lead by example.

• Lead the team responsible for developing SoC models end‑to‑end, including model architecture, integration with other model or infrastructure components, testing, and debug
• Mentor and develop the team, while hiring top talent to continue scaling
• Lead new product and feature development both at a high level and in the trenches - this is a hands‑on role, and you’ll be expected to dive into technical issues and the codebase head‑first
• Work closely with architecture, RTL design, design verification, emulation, and software teams to build, debug, and deploy your models
• Focus on customer needs and use‑cases to improve our models’ quality, delivery speed, and usefulness within AWS
• Develop mechanisms and infrastructure to enable the team to build high quality, maintainable, well documented, modular, and testable software

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 and tested with high quality. Our SoC model is critical for both the SoC development process and for production software development. 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 how we’re using modeling to build successful products.

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.

We’re changing an industry. We’re searching for individuals who are ready for this challenge, who want to reach beyond what is possible today. Come join us and build the future of machine learning!

Trainium3 () production software development leveraged our virtual platform before silicon arrived. This enabled running a full training workload within 12 hours once silicon arrived, and getting servers in customer hands in mere months. While these are amazing feats, there’s significantly more value the modeling team can provide - we need to scale, with the right leadership, to enable this. We are looking for an SDM who loves getting their hands dirty, but enjoys management and driving high‑performance teams.

• 7+ years of engineering team management experience
• 15+ years of non‑internship professional software development experience writing functional or performance models for SoCs, CPUs, GPUs, and/or ASICs
• Familiarity with SoC, CPU, GPU, and/or ASIC architecture and micro‑architecture
• Experience building large‑scale, OOP software projects in C++ and/or SystemC

• Experience designing and developing large scale, high‑traffic applications
• Experience hiring, developing and promoting engineering talent
• Experience running metrics reports, analyzing data to identify problems and using data analysis to generate and explain metrics
• Experience in computer architecture, or experience with Machine Learning and Large Language Model fundamentals, including architecture, training/inference life cycles, and optimization of model execution
• Experience building high‑performance / multi‑threaded / distributed software systems
• Experience developing and calibrating performance models for custom silicon chips
• Experience building test regression and continuous integration…