Software Engineer, Model Inference, DeepMind

Software Engineer, Model Inference, DeepMind Overview

At Google DeepMind, we are dedicated to creating the world’s first general-purpose learning agent. As a Software Engineer specializing in Model Inference, you will engage with cutting-edge AI agents developed by our talented team of Machine Learning and Neuroscience researchers. Your role will encompass a variety of tasks, including the creation of systems for agent testing using both 2D and 3D games, developing test problems in physics simulators, and producing graphical visualizations of results. You will also build competitive leaderboards for agents and test new algorithms on robotic systems.

To excel in this position, you should possess a strong foundation in software engineering and a passion for tackling diverse and challenging problems within a mission-driven team. You will play a crucial role in translating AI research into practical applications, working closely with researchers and engineers to optimize and deploy large language models (LLMs) like Gemini onto Google’s production infrastructure. This position will have a significant impact on users across various applications, requiring a blend of technical expertise and collaborative problem-solving to ensure efficiency and quality throughout the LLM deployment lifecycle.

This role offers both individual contributor and team lead opportunities, welcoming candidates from both Software Engineering and Research Engineering backgrounds. We encourage applicants with both specialized and general interests in serving to apply.

Responsibilities include:

• Collaborating with research teams to ensure next-generation modeling approaches are designed and implemented with production considerations.
• Working with infrastructure teams to deliver serving infrastructure optimized for efficiency and performance, addressing bottlenecks in speed, scale, and quality.
• Identifying opportunities to automate tasks, eliminate redundancies, and improve the overall velocity of model releases.
• Gaining a deep understanding of serving frameworks, pre-processing pipelines, caching mechanisms, and other relevant technologies.
• Leveraging roofline analysis, hardware-level profiling, and systems analysis to identify and eliminate performance bottlenecks across ML frameworks, compilers, custom kernels, and serving infrastructure on hardware accelerators.

Minimum qualifications include a Bachelor’s degree or equivalent practical experience, at least 8 years of software development experience, and a minimum of 2 years in deploying and maintaining machine learning models in a live production environment. Candidates should have experience with profiling, configuring, or executing ML workloads on hardware accelerators, as well as designing and optimizing model serving infrastructure.

Preferred qualifications include experience with developing serving infrastructure, programming hardware accelerators via ML frameworks or low-level programming models, and familiarity with distributed ML systems optimization.

Join us at Google DeepMind, where we are committed to using our technologies for public benefit and scientific discovery, ensuring safety and ethics are always our top priorities.

Degree Requirement: Not Specified

Visa Sponsorship May be