Google is seeking a Software Engineer III to advance AI/ML efficiency for YouTube in Mountain View, California. This on-site role focuses on optimizing the models powering the YouTube algorithm, with emphasis on low-precision quantization, knowledge distillation, parameter sharing, and hardware-friendly model architectures to reduce training and serving costs while maximizing fleet utilization and user value.

Responsibilities include analyzing ML workloads to identify compute and memory bottlenecks, optimizing accelerator utilization to maximize compute efficiency, implementing efficiency techniques, and productionizing algorithms for real-time training and inference. You will collaborate with ML model developers and researchers to co-design hardware-friendly architectures and deploy universal efficiency libraries, as well as optimize serving data pipelines and distributed systems.

Qualifications: Minimum: Bachelor's degree or equivalent practical experience; 2 years of experience in ML infrastructure (model deployment, evaluation, data processing, debugging) and accelerator hardware architectures (TPUs/GPUs); 2 years programming in C++ and Python; experience with ML frameworks such as TensorFlow, JAX, PyTorch, or ML compilers like XLA; experience with ML model optimization.

Preferred: Master’s degree or PhD in CS or related technical fields; experience developing accessible technologies; experience debugging correctness and performance across the ML stack; experience with ML compilers and optimization passes.

Compensation: US: $147000 - $210000 (USD) plus 15% t

Google is seeking a Software Engineer III to advance AI/ML efficiency for YouTube in Mountain View, California. This on-site role focuses on optimizing the models powering the YouTube algorithm, with emphasis on low-precision quantization, knowledge distillation, parameter sharing, and hardware-friendly model architectures to reduce training and serving costs while maximizing fleet utilization and user value.

Responsibilities include analyzing ML workloads to identify compute and memory bottlenecks, optimizing accelerator utilization to maximize compute efficiency, implementing efficiency techniques, and productionizing algorithms for real-time training and inference. You will collaborate with ML model developers and researchers to co-design hardware-friendly architectures and deploy universal efficiency libraries, as well as optimize serving data pipelines and distributed systems.

Qualifications: Minimum: Bachelor's degree or equivalent practical experience; 2 years of experience in ML infrastructure (model deployment, evaluation, data processing, debugging) and accelerator hardware architectures (TPUs/GPUs); 2 years programming in C++ and Python; experience with ML frameworks such as TensorFlow, JAX, PyTorch, or ML compilers like XLA; experience with ML model optimization.

Preferred: Master’s degree or PhD in CS or related technical fields; experience developing accessible technologies; experience debugging correctness and performance across the ML stack; experience with ML compilers and optimization passes.

Compensation: US: $147000 - $210000 (USD) plus 15% t