About AnvilEight
AnvilEight is a Python engineering company that takes responsibility for the systems it
builds — production web, automation, and AI for UK, EU, and US businesses — founded in 2010, with ongoing
client partnerships that span more than a decade.
Most of our work is on long-running client products, so the same engineers often stay
on a product for years. Most of the people who work with us have been here for five years or more.
Recent work includes a lead platform for a UK estate agency group, an AI back office for
a field-service company, an AI automation platform for Runa, and research study platforms for a US
university.
See the case studies →
The role
You work on client products that run for years, on both sides of the same feature: the Django API and the React app its users work in, such as dashboards, self-service portals and admin tools. You own features from the data model to the screen, and you talk to the client about them yourself.
You join a product team of three to seven people: engineers, QA, a
project manager, and the client's product owner, who sets the priorities. Teams work in short
sprints with planning, demos and retrospectives, in the client's own repositories and tools.
What you'll do
- Build features end to end across the stack: Django and DRF API, React and TypeScript UI, tests and deploy.
- Turn client requests into a written spec with acceptance criteria before any code is written.
- Design the API contract with both sides in mind: errors, pagination, auth, and types that match on the back end and the front end.
- Build responsive, accessible interfaces from the designer's mock-ups.
- Keep the front end fast: bundle size, rendering and data fetching.
- Write tests on both sides: pytest for the API, component and end-to-end tests for the UI.
- Review your teammates' code, and keep the project's tests, CI and rules file in good shape.
- Work with the client directly, in English: clarify requirements, estimate, demo and report risks.
What you need
Experience
- Typically 4+ years of production work with Python and React. What we check is how well you understand both sides, not the years alone.
- Features you have shipped on both sides of the same product, and then maintained.
Back-end fundamentals
- Python itself: the data model, iterators and generators, context managers, async vs threads vs processes.
- HTTP and the web: the request lifecycle, headers, cookies and sessions, CSRF, caching.
- SQL written by hand: indexes, joins, transactions, reading EXPLAIN, spotting the ORM's N+1 queries.
- Data structures and complexity in practice; a testing strategy: what to test, and what to mock.
- Linux, git, Docker and networking basics for debugging.
- Knowing what Django does for you and how, including zero-downtime migrations.
Front-end fundamentals
- The browser: the DOM, the event loop, rendering, the network tab, CORS, cookies vs tokens.
- JavaScript and TypeScript: closures, promises and async, types as design.
- React: state and effects, render cycles, data fetching and forms.
- Semantic HTML, accessibility, CSS flex and grid, responsive layout.
Ways of working
- Daily use of an AI coding assistant, with the practices described under "How we build software", and a real task you can walk us through.
- English at B2 or higher: you talk to clients in writing and on calls.
- Regular overlap with the European working day.
Nice to have
- React Native
- GraphQL
- End-to-end testing with Playwright
- Design systems and component libraries
- LLM features in the UI: streaming responses, chat interfaces
- AWS (EC2, RDS, S3)
How we build software
AI-assisted coding is part of the job
AI writes code fast. We hire engineers who can tell when it is wrong, and that takes
fundamentals. You use an AI coding assistant every day: ours is Claude Code, and Cursor, Copilot and Codex are
fine too. You know these practices and can show them:
- a rules or project-memory file (CLAUDE.md, AGENTS.md or Cursor rules) kept in the repository, so the agent
follows the project's conventions;
- spec first: acceptance criteria and named tests before the agent writes code;
- tests and CI as the gate on agent output, not the agent's own word that it is done;
- small, reviewable diffs: you review every AI-written line as your own, and then a second engineer reviews
it;
- giving the agent context through MCP (database, logs, docs) while keeping secrets and untrusted input out
of it.
Every change goes through the same gates
- SpecAcceptance criteria and named tests, approved by a person
before any code.
- BuildOne commit per task, each with its named test
passing.
- CheckThe full test suite and the required CI checks,
including secret scanning.
- ReviewAn AI review against the spec, then a second
engineer.
- ShipMerge on green CI. The deploy is watched and
verified.
At the interview, be ready to walk us through a real task you did with an AI
assistant: the rules file, the spec or prompt, the tests, and what you changed in its output.
Your first three months
What we expect, so you know what good looks like from the start.
-
Weeks 1–2
Get access, read the project's docs and rules file, run it locally, and ship a first small change to production through the full process: spec, tests, review, deploy.
-
Months 1–2
Own features on one client product across the API and the UI. Your estimates hold, and your pull requests pass review in a round or two.
-
Month 3 on
The client and the team rely on you for an area of the product: you plan the work in it, keep the API contract clean, and improve its tests and rules file.
What we offer
-
Pay
Up to $5,000 a month for full-time, depending on the interview. The hourly rate is
agreed individually.
-
Contract
A full-time contract or hourly work, whichever suits you. We work with engineers under contracts,
not employment.
-
Remote
Work from anywhere, with overlap with the European working day. You plan your own day.
-
A team that stays
Long-running products and stable teams: most of us have been here for five years or more.
-
Time off
On a full-time contract: 22 working days of paid leave a year, and paid sick days with no sick
note.
-
Learning
On a full-time contract: free English classes, technical lectures and courses.
How we hire
- ApplySend your CV and, if you like, a short cover
letter.
- CV screeningOur HR manager checks that your experience fits
the role.
- First callHR gets back to you to talk about the role, the
terms and your questions.
- Test taskA practical task, the same for every candidate for
the role.
- Technical interviewWe go through your test task and the
fundamentals, and you walk us through a real task you did with an AI assistant.
- OfferOnce we are sure we fit each other, we make you an
offer.
Every candidate for a role is assessed on the same criteria and given the same test
task.
Questions
Can I work hourly instead of full-time?
Yes. The role is open to a full-time contract or hourly work. The salary shown is for full-time; the
hourly rate is agreed individually.
Where do I need to be?
Anywhere. We need regular overlap with the European working day for calls with the team and the
client.
Is this an employment contract?
No. We work with engineers under contracts, full-time or hourly, not employment contracts.
How good does my English need to be?
B2 or higher. You talk to clients directly, in writing and on calls.
Do I need a cover letter?
No, it is optional. If you write one, tell us about a real task you did with an AI coding assistant:
the rules file, the spec or prompt, the tests, and what you changed in its output.