Graduate Software Engineer

Location: London (office-based, ~4 days per week)

Build Something That Matters

native has been building for ten years and still runs like a startup: small, fast, and unsentimental about how things get done. We run a managed marketplace that connects students, Students' Unions, universities and advertisers. We increase student engagement, help Students' Unions fund themselves properly, and give advertisers a measurable route to a student audience. The closer those three line up, the better the business works.

What We're Looking For

• You think from first principles and build answers from the ground up, not from the borrowed one.
• You can decide when there's no map, and you build structure where there isn't any.
• You care that things are done properly. That's reason enough to do them properly.
• You have range. Not just sharp on paper: you've done things that demanded resilience, judgement or initiative.
• We're open to a wide range of degrees. Intellectual sharpness and structured thinking turn up often in engineering, maths, computer science, philosophy, languages or history, but not always, and not only there. If your path is less typical, tell us how it shaped the way you think and why that stands up.

What You'll Be Working On

This is a broad build role. The work runs from the pipelines that move and model our data to the applications that put it in front of people. The mix of software engineering, data engineering, data science and analysis shifts week to week, and we expect you to move between all four. You'll be hands‑on with:

• Shipping production Python services in FastAPI, internal tools and dashboards, and front‑end work in Jinja, Tailwind and React, across Heroku and AWS.
• Building and maintaining data pipelines in dbt and BigQuery.
• The models behind our student personas: clustering and scoring students on their interaction data, labelling it (sometimes with LLMs), and turning noisy signals into something commercially useful.
• Identity stitching, so a student looks like one person across sources that don’t agree out of the box.
• Applying our pseudonymisation and data minimisation practices as you build. You won’t own this, but you’ll be trusted to get it right.
• Finding what’s slow, fragile or held together with tape, and fixing it because you were the one who noticed.

How The Work Gets Done

We build with agentic coding tools, and you will too. This is not a perk and not a line about being comfortable with AI. It’s how an engineer here ships in an afternoon what used to take a week. That raises the bar rather than lowering it. The model is fast and often wrong in ways that look right, so the job is judgement. You frame the problem and decide what a good answer looks like before you let the model near it. You treat what it gives you as a first draft to be checked, not an answer to be trusted, and you catch the version that compiles cleanly and is quietly broken. When you open a pull request, you own every line in it, including the ones you didn’t type, and you can stand behind them with the tool closed.

If that sounds like more work than writing it yourself, sometimes it is. The engineers who get the most out of these tools are the ones who were already rigorous. That rigor is what we’re hiring for.

Required Skills

• You’ve excelled at something, and we’re not precious about the form: first‑class honours, a Dean’s List, a research result, a project you couldn’t leave alone. We’re reading for rigour and clarity of thought.
• You write proper Python, not only notebook Python. At home exploring data with pandas and numpy, equally at home writing a small service someone else can run without you in the room.
• You write SQL with intent. Not just queries that return the right rows, but ones that stay clear when the data’s messier than the example.
• You’ve worked with real, messy data: designing a schema, cleaning a dataset that fought back, checking your results are actually true. Coursework, Kaggle, a personal project, wherever.
• You teach yourself the tool you need before anyone tells you to. Data side: BigQuery, dbt, Airflow, Docker. Software side: git, a web framework, getting something live on the cloud.
• Bonus points if you’ve built and shipped something end to end that other people used. A tool, an app, an API, a bot. Anything real.

Progression

This is a six‑month engagement, and we mean it as a proving ground for a permanent hire, not an internship and not a rotation. Do well and you move into a promoted, permanent role at the end of it. The trajectory is the offer here. You drop into live production from week one with real ownership, and the breadth is the point: in six months you’ll have shipped across software, data and ML. That’s rare this early, and almost impossible to get on a scheme that keeps you in one lane while it decides what to do with you. During the process you’ll talk to grads who joined this way, so you hear how it actually went from them rather than from us.

Location and ways of working

You’ll work from our London office at least four days a week, with one optional day remote. We move fast and decide fast, and most of that happens face to face.

How To Apply

• Answer a trade‑off you had to make, and how you decided.
• Answer a problem you tackled without much guidance.
• Explain a system or process you’d redesign, and how you’d go about it.
• Describe a time you chose what not to do, and why.
• Include a recent CV, or a link to your LinkedIn or equivalent.
• And if you’re reading this thinking you want it but probably won’t get picked, apply anyway. We care far more about how you think and how you show up than whether you tick every box you imagine we’re counting. Don’t rule yourself out.

We hire on a rolling basis. If this is the kind of challenge you’re ready for, get in touch.

Equal Opportunity Statement

We’re building an equitable environment where everyone at native can do the best work of their lives. Diversity and inclusion sit at the centre of that, and we put real support behind helping all of our people grow here.