Senior Staff Software Engineer, ML Performance, GPU

Google•Sunnyvale, CA

20h•$262,000 - $364,000

About The Position

Google's Core ML organization is hiring a software engineer for its ML Performance team to drive GPU performance optimizations for Gemini models. In partnership with DeepMind and XLA, the team develops custom kernels using CUDA, Triton, and Pallas/Mosaic to improve efficiency. Its work includes performance analysis—such as headroom and roofline analysis, benchmarking, and hill-climbing—and optimization techniques including sharding, compiler optimizations, quantization, custom kernels, and agentic optimizations. The team also turns model-specific improvements into scalable, stack-level enhancements across compilers and kernel libraries to improve out-of-the-box performance. The position is part of Google’s AI and Infrastructure team, which develops AI and infrastructure capabilities at scale, including platforms and computing systems that support Google products, customers, and users. Google software engineers work on critical projects and may work across different technologies and teams as business needs evolve.

Requirements

Bachelor’s degree or equivalent practical experience.
8 years of experience in software development.
7 years of experience leading technical project strategy and ML design, and optimizing end-to-end machine learning, such as model deployment, model evaluation, data processing, debugging, and fine-tuning.
Experience with modern GPU architectures, memory hierarchies, and performance bottlenecks.
Experience with low-level GPU programming, such as CUDA, Triton, or CUTLASS, and performance engineering techniques.
Experience with modern LLMs and their deployment on AI accelerators.

Nice To Haves

Master’s degree or PhD in Engineering, Computer Science, or a related technical field.
8 years of experience with data structures and algorithms.
5 years of experience in a technical leadership role leading project teams and setting technical direction.
3 years of experience working in a complex, matrixed organization involving cross-functional or cross-business projects.
Experience with machine learning algorithms and tools.

Responsibilities

Identify and maintain LLM training and serving benchmarks, use them to identify performance opportunities, drive XLA:GPU/Triton performance toward state of the art, and guide XLA releases.
Collaborate with DeepMind to address ML model performance challenges, including onboarding new LLM models and products on GPU hardware and enabling efficient large-scale LLM training and serving.
Run architecture-level simulations on GPU designs, perform roofline analysis to guide internal teams, and explore novel model and data efficiency techniques.
Analyze performance and efficiency metrics to identify bottlenecks and design and implement solutions at Google fleet-wide scale.
Collaborate with tooling and fleet metrics sub-teams to build tools that track performance and efficiency and extract metrics from Google’s running workloads.