ML Performance Engineer - GenAI/LLMs, TPU Optimizations

Google LLC; Mountain View (CA); Presencial; USD 174.000 - 252.000; Jornada completa

Ventajas ofrecidas por este puesto de trabajo

Equity

Benefits

Descripción de la vacante

Google DeepMind seeks a Senior Software Engineer, ML Performance, GenAI, in Mountain View, CA. You will work on the world’s first general-purpose learning agent, building systems for agent testing, visualization, and production ML workloads with a focus on LLM inference, latency, throughput, and cost per token.

You will collaborate with Gemini teams and TPU hardware groups to push accelerator capabilities, test new algorithms on robots, and develop scalable performance improvements for research

Formación

• Bachelor's degree in Computer Science, Software Engineering, Mathematics, or related STEM field or equivalent practical experience.
• 5 years of experience in software development, systems programming, or applied ML engineering or equivalent education credit.
• 3 years of experience with performance optimization, systems data analysis, visualization tools, or debugging.
• 3 years of experience with Generative AI, Large Language Models (LLMs), Machine Learning and related frameworks.
• 1 year of experience designing and launching production software products or foundation AI systems.

Responsabilidades

• Drive down the cost per token of Gemini inference by optimizing throughput, accelerator utilization, batching, scheduling, and KV cache efficiency.
• Be among the first to bring Gemini to new TPU generations by characterizing hardware at launch, tuning kernels, sharding, and serving configurations for memory hierarchies.
• Implement model compression techniques—such as quantization and sparsity—to reduce cost-per-token for production Gemini models.

Conocimientos

Performance optimization

Generative AI

LLMs

Production software

Agentic workflows

Educación

Bachelor's degree

Herramientas

PyTorch

JAX

Triton

vLLM

Gemini API

Descripción del empleo

Google DeepMind seeks a Senior Software Engineer, ML Performance, GenAI, in Mountain View, CA. You will work on the world’s first general-purpose learning agent, building systems for agent testing, visualization, and production ML workloads with a focus on LLM inference, latency, throughput, and cost per token.

You will collaborate with Gemini teams and TPU hardware groups to push accelerator capabilities, test new algorithms on robots, and develop scalable performance improvements for research