About the role
Arva AI is building the AI Agent platform that compliance teams at banks, fintechs, and crypto firms use to automate sanctions screening, KYC/KYB, adverse media, and AML transaction monitoring. We process over a million reviews a month and our agents resolve the vast majority of them without human intervention.
As a Senior Full-Stack Engineer you'll take end-to-end ownership of product features — from the React interface a compliance analyst uses to the Node.js services and AI pipelines running underneath it. You'll work on capabilities like document fraud detection and web-based due diligence, and you'll build the feedback loops — prompt optimisation, fine-tuning, custom model training — that make our agents measurably better over time. This is a five-days-in-office role at our Central London office, reporting into the engineering team.
What you'll do
- Design, build, and ship full-stack product features across front-end, back-end, and AI layers, taking each from brief to production.
- Develop and improve AI systems including computer vision and multimodal models, with a focus on document fraud detection and deep web due diligence.
- Build continuous-improvement pipelines: prompt engineering, fine-tuning pre-trained models, and training bespoke models for specific compliance use cases.
- Own reliability and scalability — instrument your systems, catch problems early, and close the loop so they don't recur.
- Collaborate with product, research, and compliance domain experts to turn complex regulatory requirements into working software.
Who you are
You have at least four years of full-stack product engineering experience, with a clear record of shipping production features across both front-end and back-end. You're fluent in TypeScript, React, and Node.js and you know what it takes to build systems that hold up under real load.
Critically, you've worked hands-on with LLMs — prompt engineering, fine-tuning, and deploying language or vision models in production — not just as a user but as an engineer responsible for their behaviour.
You take genuine ownership: when something breaks or a requirement shifts, you drive it to resolution rather than waiting to be directed.
Nice to have: experience training custom models from scratch for narrow use cases; familiarity with CI/CD, DevOps, and MLOps tooling; and a track record of delivering in early-stage environments where the roadmap moves fast.
Why join
Arva is backed by Y Combinator and Google's AI Fund and is at the stage where individual engineers shape the product and the architecture. You'll work on a genuinely hard problem — financial crime compliance is high-stakes, data-rich, and underserved by good software — with a team that moves quickly and takes quality seriously. Compensation includes equity, and salaries are reviewed every six months. Relocation support is available.