Software Engineer, TPU Compiler, PhD, Early Careers

Software Engineer, TPU Compiler, PhD, Early Careers Overview

This role sits within Google’s software engineering organization and focuses on the TPU compiler work that underpins large-scale machine learning execution. The team builds the Accelerated Linear Algebra compiler stack used to split, optimize, and run demanding ML models across multiple TPU accelerators, supporting both internal users such as Google DeepMind and external Google Cloud customers. The work is closely tied to the infrastructure behind Google Gemini and to Google Cloud’s broader mission of helping organizations transform digitally with advanced, enterprise-grade technology.

As a software engineer in this position, you would be expected to contribute to a critical project while also remaining adaptable enough to move across teams and initiatives as the business changes. Google describes the engineering environment as one that values versatility, leadership, and a willingness to take on new technical problems across the stack. The role is based in London, UK, and is presented at a mid-level experience band.

What you would do

• Build and maintain TPU compiler functionality in C++, contributing directly to product and system development work.
• Join design discussions with peers and stakeholders, and sometimes take the lead in those reviews to compare technical options and settle on the best approach.
• Help advance a compiler that can scale machine learning models across accelerator hardware such as TPUs and GPUs for use across Google and Google Cloud.
• Develop optimizations and important features that improve performance and help key production teams move faster.
• Apply artificial intelligence methods to both the process of building the compiler and to the compiler technology itself.

What Google is looking for

• Strong programming ability with a solid foundation in data structures, algorithms, and software design.
• Research experience in a relevant technical area, which may include AI, distributed systems, machine learning, data mining, natural language processing, image classification, spam fighting, or a similar discipline.
• Hands-on or academic exposure to machine learning or artificial intelligence.
• Either current enrollment in a PhD program or completion of one.
• Experience working with parallel computing systems.
• Background in compiler development and compiler construction.
• Ability to debug complex systems effectively and write software for concurrent or parallel workloads.
• Experience with accelerator platforms such as VLIW systems, vector machines, GPUs, or DSPs.
• For this level, the expectation is that you can make progress independently, solve difficult problems, and support more junior engineers through mentoring.

Additional context about the team and role

The TPU team’s compiler work is a core part of Google’s machine learning infrastructure. It is responsible for partitioning and optimizing workloads so they can run efficiently across accelerator hardware. Because the compiler supports both internal research and production use cases, the work has direct impact on major systems and customer-facing cloud capabilities.

Google Cloud’s platform serves organizations in more than 200 countries and territories, and the company positions this role as part of that broader effort to provide scalable, sustainable, and high-performance technology. The engineering culture emphasizes broad technical curiosity, the ability to work across different problem domains, and readiness to contribute to fast-moving projects.

Other information included on the page

Google states that it is an equal opportunity and affirmative action employer and says it is committed to building a workforce that reflects the people it serves, fostering belonging, and providing fair employment opportunities regardless of protected characteristics. The page also notes that applicants who need an accommodation can use Google’s accommodations process.

The posting says English proficiency is required for all roles unless the job ad explicitly says otherwise, because Google operates globally and relies on effective collaboration and communication across regions. It also includes standard recruiting notices stating that Google does not accept agency resumes, does not want resumes forwarded to its jobs alias or employees, and is not responsible for fees tied to unsolicited resumes. The page further explains that any equity grants are handled by Alphabet Inc. under the terms of the relevant stock plan and grant agreement, and that such grants are discretionary and not guaranteed.

To apply, candidates should use the Google Careers application process shown on the posting page.

Degree Requirement: PhD

Visa Sponsorship Promising