Staff Engineer (Full-Stack, Infra & AI)

Run an entire production platform — and the AI on top of it. Report to the executive team. Decide what matters and ship it. Most senior engineering jobs are a seat at a big table: a slice of a service, a queue of tickets, three layers of sign-off before anything ships. This is the opposite.

We have a live product that real customers depend on every day, and we're looking for one exceptional engineer to own it outright — the backend, the web app, the infrastructure underneath, and the LLM‑powered features customers actually use. You'll have the keys, a direct line to leadership, and the trust to run it your way. If you've ever thought "I could move so much faster if you just let me own this end-to-end" — this is that role.

About BXBX is an AI-powered farm intelligence platform. We work with the largest perennial crop operations in the world and give them something they have never had: a single, reliable view of their own business. The opportunity is enormous. Agriculture loses roughly $500bn a year to data that gets collected and never used. A grower running thousands of acres will have that data spread across a dozen systems that do not connect, so the people who should be making decisions spend their weeks on admin, audits and chasing information. Most farm software has made this worse, not better.

We built BX to fix it. Our technology connects to any system a farm already uses and unifies the data automatically. Our agentic system, SAM AI turns that data into insights, compliance reports and financial intelligence, replacing manual work that costs a large operation between £50,000 and £200,000 every year. It is working. We are live with major growers across Washington State, California, Spain and New Zealand, covering more than 80,000 acres, and we're growing.

Now we are going deeper. More growers, more regions, and a product that moves from telling growers what happened to telling them what to do next: field planning, irrigation and labour forecasting, grower-to-packhouse visibility. BX will become the system large farming enterprises run on around the world.

Why this is interesting engineering. The data is genuinely hard. It arrives from sensors, handheld scanners, spreadsheets emailed at midnight, scanned PDFs and packhouse systems built decades ago, and all of it has to reconcile into something a grower will bet money on. If you want problems where the messy real world fights back, there is no shortage here.

Why Is This a Rare Seat? Total ownership. You're not one of ten engineers on the platform — you are the platform engineer. What ships, how it's built, what gets fixed first: your call. A direct line to the top. You report into the C-suite. No middle management, no ticket factory, no waiting on a committee. Real AI work, not an AI job title. You'll own LLM features that are already in production in front of customers — agent flows, tool‑calling over real customer data, evals that decide what ships. Very few roles let one person own both the platform and the intelligence layer on it. Immediate, visible impact. Everything you do is in production in front of customers — no work that dies in a backlog or a reorg. Autonomy that's actually real. You set priorities against what leadership cares about, then execute without anyone looking over your shoulder. Modern stack, modern tools. TypeScript top to bottom, a clean cloud setup, and AI harness tooling (Claude Code and friends) as a first-class part of how we work — not something you have to sneak past anyone. Room to grow with the company. As the team scales and senior hires come on board, you'll already be the person who knows the platform cold — the natural foundation everything else is built on.

What You’ll Own The whole application platform — a NestJS backend, a Next.js web app, and the GCP infrastructure that runs them. It's yours to understand, shape, and keep healthy. The AI features on top of it — an in-product assistant that answers customer questions against their own data, and LLM‑driven extraction that turns messy documents into structured records. You own the prompts, the tools, the model choices, and the evals that prove a change was an improvement. New customer data, end-to-end — when a customer says "we need X," you take it from idea to shipped: model it, build the ingestion, wire up the integration, surface it in the product. Reliability and craft — you keep the platform fast, stable, and maintainable. You'll fix what breaks as readily as you build what's next; great owners do both without drama. You’ll work alongside a Senior Data Engineer who owns the data pipeline and data modelling. Everything the product runs on top of that is your domain.

The Scope This is a lot of ground for one engineer, and we're not going to pretend otherwise. It was previously covered by more people. You'd be taking it on as the single owner — with a data engineer beside you, leadership directly reachable, and more senior hires expected as the company's new strategy lands. That's the trade: more surface than is comfortable, in exchange for more ownership than you will realistically be offered anywhere else at this level. If you'd want the whole thing on your plate, that's a feature. If you'd want it narrower, this isn't the right role — and we'd both rather know that now.

Who Thrives Here The engineer who wants this is the one who's tired of asking permission. You like having the whole board in front of you. You'd rather own a real product with real customers than prototype something that may never ship. You're calm when things break, decisive about what matters, and you get a genuine kick out of being the person everyone trusts to keep the thing running and make it better. If you need a big team around you, tightly-scoped tasks, or a well-worn playbook, this won't be your favourite job. If you want the keys, it's hard to beat.

What We’re Looking For

• 6–8 years building and running production web apps, with real examples of owning systems end-to-end — not just contributing features.
• Strong across the stack: TypeScript / Node (NestJS ideal) on the backend and React / Next.js on the frontend. Genuine full-stack range matters here — there's no one to hand the other half to.
• You’ve shipped LLM features to production — prompts, tool calling against real data, and some honest way of telling whether the output got better. Owning a feature customers used counts; training models is not required.
• Confident with PostgreSQL (schema design, migrations, query debugging) and comfortable operating cloud infrastructure (GCP ideal; AWS/Azure fine).
• A track record of owning reliability — you’ve chased incidents across services and you reach for logs before guesses.
• Self-sufficient and high-trust — ramps fast on an unfamiliar codebase, thrives on autonomy, and stays calm in startup ambiguity.
• Bonus points: Python, Terraform, agent frameworks or RAG beyond single-prompt features, LLM tracing and cost control, hands‑on AI harness tooling (Claude Code), and any experience with messy real-world data — agriculture, climate, sustainability, logistics, industrial operations. If you've worked somewhere the data came from the physical world rather than a web form, you'll recognise the problems here.

Your First Six Months

• Month 1: You've got the platform in your head — UI to API to database to infra — and you're already shipping fixes to production.
• Months 2–3: You're the de facto owner: incidents handled independently, enhancements shipping, your first customer data request delivered end-to-end, and your first improvement to an AI feature out with evidence it helped.
• Months 4–6: You run the platform and its AI features with full autonomy, the backlog is visibly shorter, the system is more solid than you found it — and you're ready to take on bigger things as the company grows around you.

Practical Details Location: London area. Hybrid, one day a week in the London office. Compensation: £100,000–£110,000, plus participation in the company equity scheme. Reporting: Into the executive team.

Everyone belongs at BX We celebrate all genders, identities, races, ethnicities, sexual orientations, ages and backgrounds, and none of these factors impact how we review job applications. We care a lot about creating a great culture. It’s something we’re always working on. We make sure everyone who joins our team feels comfortable in the environment we create, feels valued by everyone they work with, and feels excited about being part of the team.