Software Engineer, ML/AI Reference Models

In this role, you'll work to shape the future of AI/ML hardware acceleration, focusing on Google's TPU (Tensor Processing Unit) technology. You'll be part of a diverse team developing custom silicon solutions for Google's most demanding AI/ML applications. Your responsibilities will include developing functional and/or performance models for ML compute IPs, integrating them with the Cloud TPU SoC model, and working closely with ML and SoC architecture teams. You'll collaborate with pre-silicon verification, post-silicon validation, and software teams to implement these models in their validation flows and aid in delivering high-quality designs for next-generation data center accelerators.

Key responsibilities:

• Study ML IP instruction sets and architectures, driving discussions on new features
• Develop functional and/or performance models for ML IPs
• Integrate functional models with the Cloud TPU SoC model
• Work with verification and validation teams to deploy models in their flows
• Collaborate with compiler and software teams to enable early development

This role is part of Google's Technical Infrastructure team, which builds and maintains the architecture behind Google's product portfolio. You'll contribute to the innovation powering products used by millions worldwide, leveraging your expertise in complex digital design verification with a focus on TPU architecture and its integration within AI/ML-driven systems.

Qualifications:

• Bachelor's degree in Computer Science or related field (Master's preferred)
• 2+ years of experience developing simulators or reference models for hardware
• Proficiency in C/C++ and scripting languages (e.g., Python/Perl)
• Experience in modeling Machine Learning IPs, specifically compute IPs
• Familiarity with industry-standard AI/ML topologies (e.g., RESNET, BERT)
• Experience integrating IP models with SoC models

Join Google's team and contribute to cutting-edge AI/ML hardware acceleration technology!