AI Engineering Intern

The role

You will build systems that identify mistakes, improve their own behaviour, and prove those improvements work. You will work on agents, models, and infrastructure to make organic software reliable.

What you will do

• Help agents learn from their work by shaping the process of updating repository-specific instructions.
• Train and deploy LLM and ML models.
• Fine-tune SoA LLM models.
• Run experiments and compare results against the existing pipeline.
• Leverage ML tools to enhance the ability of Agents.
• Take full end to end ownership of an agent.
• Build evaluations and review tools.
• Define what a correct answer looks like and create tests.
• Scope problems, build solutions, debug regressions, and follow through until delivery.

Required qualifications

• Currently studying at University of Oxford, or University of Cambridge.
• A silver medal or higher at the International Mathematical Olympiad (IMO) or International Olympiad in Informatics (IOI), or an equivalent level of achievement in another highly competitive endeavour.
• Strong programming ability.
• Something you’ve built beyond coursework or tutorials: a model, agent, tool, or application that does useful work.
• Hands-on experience with AI tooling, such as LLM APIs, agent frameworks, model fine-tuning, or automated workflows.
• A habit of investigating why a system failed and testing whether your fix addresses the cause.
• Clear thinking and the ability to explain your decisions, including what you tried and why it didn’t work.

Nice to have

• Open-source contributions
• deployed models
• experience building evaluation pipelines

Benefits

• Internship compensation above the standard market rate.
• Ownership from your first day, with a path to shipping work into production.

Before you apply

• If applying with an equivalent achievement, explain the competition or field, your result, and the standard required to achieve it.
• Send your CV.
• Include a link to something you’ve shipped including what it does, which parts you built yourself, where it failed, and how you fixed it.
• Explain why you want to work on organic software at Light.

About the company

Light is building organic software that heals itself, adapts to the user and the business, and learns as it is used. Companies including Legora, Lovable, and Fuse Energy already run on Light.