G

Google

Runs Search and YouTube and builds Gemini AI

Tech Lead – Gemini Inference Performance

Engineering LeadershipConfirmed live in the last 24 hours
Full-timeLondon, UKNo salary listedIn-person

About the Job

Requirements

  • A bachelor's degree or equivalent practical experience.
  • Eight years of experience in software development.
  • Five years of experience in a people management, supervision, or team leadership role.
  • Five years of experience in a technical leadership role and overseeing projects.

Responsibilities

  • Set the technical roadmap for a team of performance engineers, support and develop direct reports, and drive prioritization across competing optimization opportunities, focusing on the highest-leverage problems.
  • Use roofline analysis, hardware-level profiling, and systems analysis to identify and eliminate performance bottlenecks across machine-learning frameworks, compilers, custom kernels, serving infrastructure, and hardware accelerators.
  • Apply a first-principles understanding of Transformer and Mixture-of-Experts model components to identify and prioritize opportunities to optimize execution efficiency and memory footprint.
  • Collaborate with research teams early in the development lifecycle to evaluate inference implications and model how architectural choices affect latency, memory footprint, and serving costs.
  • Guide and contribute to the development of custom kernels and serving optimizations.
  • Conduct deep performance profiling using hardware tracing tools to analyze accelerator utilization, memory bandwidth saturation, and interconnect latency across large-scale topologies.

Desired Qualifications

  • A master's degree or PhD in Engineering, Computer Science, or a related technical field.
  • Five years of experience working in a complex, matrixed organization.

About the Company

Google

Runs Search and YouTube and builds Gemini AI

Google runs the world's most used search engine and free services such as YouTube, Android, and Gmail, and it builds the Gemini AI models that work inside them. In 2017, its researchers invented the transformer, the neural network design behind ChatGPT and most other AI chatbots. People use these services without paying, and Google earns most of its money from ads placed beside search results and YouTube videos. Businesses rent computing power and Gemini models through Google Cloud, and Google also sells subscriptions and Pixel phones. Larry Page and Sergey Brin founded Google in 1998 as Stanford PhD students; it is based in Mountain View, California, United States, and is the main business of parent company Alphabet.

Company size
10,001+
Company stage
IPO
Headquarters
Mountain View, CA
Founded
1998

Recently Posted Jobs

Explore more opportunities at Google

See more jobs