Engineering Software Engineer expect in office

London (4 days in office expected) Full-time

Telluvian is a London-based, VC-backed pre-seed startup. We use mechanistic interpretability, our understanding of what's happening inside neural networks, to make AI cheaper and more secure. Learn more .

We're looking for a graduate backend engineer to build and run the infrastructure our research and product depend on. This is primarily an engineering role rather than a research one but you'll be working next to interpretability researchers, and you'll need to understand what they're doing well enough to know how it should shape the systems you build.

You'll join a team of fewer than ten people, and you'll own real infrastructure early. If you like large, messy, half-defined problems and get satisfaction from making things work, this will suit you.

This is a full time position with flexible working hours. 4 days a week in the London office is expected.

Start date is rolling.

• Cloud infrastructure and resource management. Configuring compute, minimising costs, keeping things up.
• Inference management for our ML systems. Getting models served reliably and efficiently. This is the hardest part of the role and the part we expect you to learn on the job.
• Hosting and maintaining our website. Occasional minor changes; this is not a frontend role.
• General backend product engineering across the stack as we grow.

• A computer science or related STEM degree, ideally with at least one relevant internship or prior work experience.
• Strong general software engineering skills. This means you can reason through a system design question and explain and/or justify your choices.
• A track record of shipping: work you've built, deployed, and kept running, whether professionally, at university, or on your own time.
• Working familiarity with ML concepts. We aren’t looking for you to do novel research, but you should understand roughly what a model is doing and why it matters for the system around it.
• Self-motivated. You'll often be working independently and you’ll need to work effectively in a self-directed way.

• Cloud platform experience (AWS, GCP, Azure)
• Dev Ops: CI/CD, containers, infrastructure as code
• Python, particularly with Num Py, PyTorch, or Transformers
• Any exposure to serving or scaling ML models

• Ownership of key infrastructure from day one.
• Meaningful equity in the company.
• Close proximity to frontier interpretability research, and the chance to grow into it if you want to.
• 28 days paid annual leave.
• Workplace pension with employer contributions.

We unfortunately don’t have the capacity to provide visa sponsorship at this time.