Making financial crime harder to commit and easier to catch
Arva builds evaluated, explainable, auditable AI agents that help financial institutions fight financial crime — pairing deep compliance expertise with frontier AI research to make anti-financial-crime operations faster and deeper.
Our Mission
On a mission to rethink financial crime compliance
What drives us
Anti-financial-crime work is some of the most important work in finance — and some of the least well served by technology. Our mission is to make it faster and deeper: AI agents that automate up to 92% of manual compliance review while strengthening, not weakening, the controls behind it. We're backed by Gradient, Google's AI fund, and Y Combinator, and already work with leading banks and fintechs across the UK and US.
How we work
Arva pairs financial crime domain experts with AI researchers and engineers. Together we build agents for Know Your Business (KYB), screening, and transaction monitoring that cut through the noise, surface what matters instantly, and leave a complete audit trail behind every decision — so compliance teams can decide with confidence, at speed.
What we believe
Fighting financial crime shouldn't be reactive or slow. It should be intelligent, contextual, and auditable — the best of machine speed and human judgment, working together.
Why this problem
Why financial crime?
Because it's one of the largest, least-solved problems in the global economy — and because the stakes are human. Money laundering is what makes fraud, trafficking, and organised crime pay. Stopping the money is how you stop the harm.
of global GDP laundered each year
of illicit financial flows ever intercepted
of screening alerts are false positives
spent globally on financial crime compliance annually
The numbers are stark. The UN estimates that 2–5% of global GDP is laundered every year — as much as $2 trillion — and only a tiny fraction of it is ever intercepted. Meanwhile, the institutions defending the system spend hundreds of billions annually on compliance, and their analysts still drown in alerts: the overwhelming majority of screening hits are false positives that consume hours of skilled human attention and return nothing.
That asymmetry is the point. Criminal networks already operate at machine speed — automated, adaptive, borderless. The teams defending against them have been left with checklists, swivel-chair reviews, and copy-paste. We started Arva because the people protecting the financial system deserve better tools than the people attacking it.
"The people defending the financial system deserve better tools than the people attacking it."
The team
Built by people who've lived the problem
Arva was founded in 2024 by Rhim Shah and Oli Wales, and brings together operators, compliance practitioners, and engineers from fintech, banking, and AI research. We work from London and New York, close to the customers and regulators we build for.
Advisors
Advised by people who've run compliance at global scale
Our advisors have chaired the UK's banking industry body, run risk at one of America's largest banks, and led financial crime at some of the world's biggest fintechs. They keep our roadmap honest and our standards regulator-grade.
backed by
Backed by investors who bet on hard problems
Arva is backed by Gradient, Google's early-stage AI fund, and Y Combinator, alongside Amino Capital, Olive Tree Capital, and angel investors from leading fintechs. Our seed round, led by Gradient, is funding the research and deployment work it takes to put trustworthy AI agents inside regulated institutions.
how we build
An applied research lab, not a feature factory
AI in compliance only works if it can be trusted — by analysts, by MLROs, and by regulators. So we run Arva like an applied research lab: every agent we ship is evaluated before it's deployed, explainable while it runs, and auditable after the fact.
Our security and AI-risk programmes are documented openly — see our Security Center and AI Risk Management.
Evals before deployment
No agent reaches a customer until it has been graded against expert-labelled datasets and adversarial cases — and every change is regression-tested against them. If we can't measure it, we don't ship it.
Explainable by requirement
Every decision an Arva agent makes comes with its reasoning, its evidence, and its confidence — written for a compliance officer, not a data scientist. Explainability isn't a feature we added; it's a constraint we design under.
Auditable end to end
Versioned models, logged decisions, documented methodology. Our agents are built to sit inside the model risk management frameworks banks already trust — and to withstand the scrutiny of an auditor or a regulator.
Forward-deployed with customers
Our engineers work directly with the compliance teams who use Arva. Small teams, full ownership, weekly shipping — and the feedback loop of watching an investigator use what you built yesterday.
our values
How we operate
01
Own With Urgency
Speed with ownership is our superpower. We value action over perfection — we get our hands dirty, experiment, iterate fast, and own the outcome completely.
02
Customer First, Team Second
Our first priority is creating value for customers. Our second is building a team people genuinely grow in. Ego comes last — collective success is the point.
03
Transparency
Open communication and full visibility, across teams and roles. Decisions, successes, and failures are shared in the open — because trust compounds.
Careers
Become an Arvanaut
Arva /ärvə/
1. "Fast" in Sanskrit — carrying connotations of versatility and power.
2. "Strength" in Arabic — symbolising stability and resilience.
We're a small team taking on one of the biggest problems in finance — and we hire people who want the responsibility that comes with that. Arvanauts get real ownership from day one, sit beside the founders and our customers, and ship work that matters in weeks, not quarters.
If you've ever wanted your engineering, research, or operations work to make crime measurably harder — this is the place to do it.
Ownership from day one
Real problems, real accountability, no waiting your turn.
London & New York
Hybrid working setups supporting our core offices.
Proximity
Work directly with founders, customers, and the compliance experts we build for.
Competitive package
Salary, equity, and benefits crafted to reward long-term value creation.
The latest
News & Research
Research notes, product updates, and company developments — straight from the team.
As featured in →
Forbes
Bloomberg
TechCrunch
VentureBeat
Writing about Arva? Access high-resolution founder photos, brand marks, and our company boilerplate in our Media Kit. For press inquiries: press@arva.ai
