AI & Automation Specialist in London

Checkout.com’s Process Architecture pillar designs how operations should work. This role builds what that looks like in practice. You’ll use modern AI‑assisted development tools (Cursor, Claude Code, n8n, LLM APIs) to rapidly prototype and ship internal tools that eliminate manual effort across our operational teams. Your primary customers are Disputes, Due Diligence, and Transaction Monitoring: high‑volume, high‑stakes teams where the right automation meaningfully reduces analyst burden and improves decision quality.

You’ll move fast. You’ll take processes that currently live in spreadsheets and manual case queues and turn them into something that mostly runs itself. Process Architects redesign how work flows; you make the new flow real.

What You’ll Do

• Use AI‑assisted development (Cursor, Claude Code) to prototype and ship tools for operational teams - fast.
• Build workflow automations that connect case management systems, document stores, APIs, and LLMs into end‑to‑end pipelines.
• Create analyst‑facing interfaces, triage tools, summary generators using self‑built applications, Retool, or similar.
• Apply LLMs where they reduce cognitive load: classification, case summarisation, structured extraction, decision support.
• Embed with Disputes, CDD, Periodic Review, and Transaction Monitoring to learn how analysts actually work - not how the process doc says they do.
• Translate vague pain points (e.g., “this part takes forever”) into shipped tools in days or weeks, not quarters.
• Demo early, iterate based on what analysts actually use versus ignore.
• Treat every tool as a product: instrument usage, capture feedback, measure time saved or quality gained.
• Document what you build and share reusable patterns so good ideas compound.

What We’re Looking For

We care more about what you’ve built and how you think than years of experience or formal credentials.

Must‑haves

• You’ve used Cursor, Claude Code, Copilot, or similar to build something real, a script, a tool, a side project, that solved an actual problem. Not just tutorials.
• Comfortable in Python, or willing to be within weeks.
• Basic API fluency: you can read a doc, make a request, parse the response, and chain a few calls.
• Familiarity with at least one LLM API. You’ve written a prompt, gotten structured output back, and understood why it sometimes goes wrong.
• Bias to ship - you’d rather get a rough version in front of users on Friday than a polished version in three weeks.
• Curiosity. You’re the kind of person who opens the hood – you want to understand how things work, not just use them.

Nice‑to‑haves

• Workflow automation tools (Zapier, n8n, Make).
• Retool, Gradio, or similar lightweight UI frameworks.
• Git, basic SQL, or comfort with JSON/CSV at scale.
• Exposure to payments, fraud, compliance, or risk operations – even tangentially.
• A portfolio, GitHub, or demo you can walk us through.