Software Engineer Summer Intern, AI Compute Libraries & Performance 2027

Description

This 12-week summer internship starts in June 2027 and gives you hands-on work on real projects alongside the team.

You will join the AI Compute Libraries & Performance team, which builds high-performance compute libraries that help customers get more from the company's hardware and help AI models run efficiently on next-generation chips.

The company builds next generation AI compute, from semiconductors and software to datacenter-scale systems. It is owned by SoftBank Group and works with sister companies in the same group.

The team combines machine learning, numerical computing and accelerator performance, and colleagues share expertise on engineering problems.

Projects are clearly scoped, and you are guided by engineers in performance and low-level software. The team uses profiling, experiments and technical reviews to compare approaches and weigh trade-offs.

A buddy helps you settle in and learn the tools and ways of working. Your manager meets you regularly to set goals, give feedback and discuss your development, including time for learning.

Questions and alternative ideas are welcome, and support grows with your responsibility.

Over the 12 weeks you will build performance engineering skills through profiling, experiments and numerical validation, and learn to explain your findings.

Responsibilities

• Write and refine C++ compute kernels for machine learning workloads, covering matrix multiplication (GEMM), convolutions, reductions, elementwise operations and fused kernels that combine several operations.
• Use profiling and microbenchmarks to measure execution speed, and study how threading, cache locality, memory layout and kernel launch efficiency affect performance.
• Build regression tests and numerical checks, investigate issues, and improve the quality of the libraries customers rely on.
• Connect numerical computing with how code runs on hardware, so the code is quick, precise and dependable.

Requirements

• You are currently working toward a bachelor's or master's degree in computer science, maths, machine learning, data science or a related field.
• You have experience with C or C++11, and you are comfortable with Python or scripting for automation and testing.
• You understand linear algebra, numerical methods or scientific computing.
• You solve problems well and work collaboratively when challenges are unfamiliar.
• Industry experience is not needed. Describe something you built or investigated through coursework, personal or team projects or other experience, how you approached it and what you learned.
• Useful extras, not essential: coursework or experience with machine learning frameworks, parallel programming, code optimisation, maths libraries or performance analysis.

Conditions

Start date: June 2027, for a 12-week summer internship.

Application deadline: midnight, 29 November 2026. Applications may close before this date, so applying early is encouraged.

Outcomes are sent by early December at the latest. Successful candidates are invited to a technical assessment, and those who reach the final stage attend in-person assessments in early spring.

You must already have the right to work in the UK for the full programme. Visa sponsorship is not available for Early Careers roles.

The company is an equal opportunity employer, and you can ask for adjustments during the interview process.

Benefits

• Flexible working to balance the internship with life outside work.
• 5 days of annual leave across the 12-week internship.
• Office life with free food, a wide range of ice cream and an on-site barista, in spaces designed for collaboration.
• Access to O'Reilly, an online learning platform, to build skills and explore interests.