About us
We are a London-based, VC-backed pre-seed startup working with mechanistic interpretability to provide model security. Telluvian — AI you can trust.
The role
This is a hands-on internship for someone early in their machine learning journey who wants real research and engineering experience inside an early-stage startup. You'll work closely with our small team (up to 10 people) on a focused interpretability project, with the opportunity to develop something real and understand what building a company from the ground up actually looks like.
The working hours are flexible, but you will be expected to work from the London office at least 4 days a week. The start date is flexible — we have mid-July in mind.
What you'll do
- Assist with machine learning research, specifically on language model interpretability
- Support our ML tooling and inference pipelines
- Software product development
What we're looking for
- Currently a student or recent graduate in ML, computer science, maths, or a related field
- Comfortable writing Python
- Some exposure to machine learning through coursework, personal projects, or self-study
- Curious, self-motivated, and happy working in a small, fast-paced team
Nice-to-haves
- Experience with standard Python frameworks — NumPy, PyTorch, Transformers
- No requirement to have particular experience in model interpretability research
- Eagerness to start own projects
What we offer
- Real research experience on a genuinely interesting problem
- Direct mentorship from people doing interpretability research day to day
- A front-row view of how an early-stage startup is built
- A meaningful, paid project you can talk about and show off afterwards
- The possibility of a return or full-time offer
How to apply
Please submit a CV and cover letter to Careers@telluvian.ai. Following the application stage there will be an initial screening interview followed by two technical interviews.
As part of your email, please include:
- Available start date
- Right to work in the UK
- Do you have hands-on experience with language model interpretability research?