Who Arva AI Hires
Arva AI sits at an unusual intersection: a research lab inside a commercial company, backed by Google and Y Combinator, teaching autonomous agents to make decisions that banks currently trust only to humans.
The company is small (20 people as of October 2026), but its hiring spans four distinct functions. Since its S24 batch, Arva AI has posted eight roles on Zero G Talent, showing a team built around product engineering and research, with commercial and data-operations layers rounding out the organization. Founders Rhim Shah (CEO) and Oli Wales (CTO) built the early team on a simple premise: banking's highest-risk decisions — AML, KYB, KYC — can move beyond human-in-the-loop review if the models are rigorous enough.
The hiring bar is explicit: Arva AI screens for operators who have shipped in constrained, high-stakes environments and who demonstrate the company's four operating values: Excellence in Everything, Agent Reasoning, Continuous Optimisation, Zoom Out Then In.
| Function | Roles posted | Example titles | Typical experience floor |
|---|---|---|---|
| Research & AI | 2 | Head of Research, AI Research Engineer | 3–6+ years |
| Product Engineering | 3 | Senior Full-Stack Product Engineer, Forward Deployed Engineer, Head of Product | 3+ years |
| Commercial / Go-to-market | 2 | Enterprise Account Executive, Strategic Customer Success | 3+ years |
| Data Operations | 1 | Data Associate (AI Labelling & Evaluation) | 1+ years |
Research and AI roles demand the most experience. The Head of Research requires six-plus years; the AI Research Engineer requires three-plus. Both sit in London, where the company's research hub has formed. Product engineering matches that bar: the Senior Full-Stack Product Engineer, Forward Deployed Engineer, and Head of Product each require three-plus years. The forward-deployed role puts engineers directly in customer environments, translating compliance workflows into product requirements rather than shipping features in isolation.
Commercial roles follow suit. The Enterprise Account Executive (New York) and Strategic Customer Success (London) both require three-plus years. Only the Data Associate (AI Labelling & Evaluation) asks for less: one-plus years, feeding the continuous-learning infrastructure the company emphasizes.
Arva AI Pay Bands
Arva AI's compensation structure shows in the roles posted to the Zero G Talent board: eight salaried positions spanning research, engineering, product, sales, and operations, with a board-wide band of $58K–$232K and a median of $152K. The careers page describes a "competitive salary and equity package, with bi-annual salary review and yearly performance based equity refresh," and adds a benefit: four weeks per year working from anywhere in the world, on top of regular team off-sites.
| Role | Location | Salary Band (annual) |
|---|---|---|
| Head of Research | London, UK | GBP 130K–180K |
| Enterprise Account Executive | New York, US | USD 150K–230K |
| Head of Product | London, UK | GBP 90K–130K |
| AI Research Engineer | London, UK | GBP 80K–120K |
| Forward Deployed Engineer | London, UK | GBP 80K–110K |
| Senior Full-Stack Product Engineer | London, UK | GBP 70K–110K |
| Strategic Customer Success | London, UK | GBP 50K–75K |
| Data Associate (AI Labelling & Evaluation) | London, UK | GBP 30K–45K |
The spread reflects a London-heavy team with a single New York go-to-market hire. At about 1.3 dollars per pound, the top London bands (Head of Research at GBP 180K (about $234K) and Head of Product at GBP 130K (about $169K)) approach the New York ceiling. The median matches the board's $152K.
Equity mechanics don't appear in the postings, but the careers page suggests a standard early-stage model: meaningful initial grants with annual performance-based refreshers. Benefits beyond the "work from anywhere" month include the bi-annual salary review cycle and team off-sites. Cole Draghi (Enterprise Account Executive, joined October 2026) and Tom Green (Senior Product Engineer, joined October 2026) both joined under this structure.
Cash bands respect London and New York markets. Equity scales with the company's banking-AI traction: Blockchain.com is a named customer.
Inside the Interview Loop
The careers page (served via Greenhouse) invites job alerts but does not publish a stage-by-stage playbook. The roles on Zero G Talent's board and their patterns imply the process.
The geographic split (London-heavy with a single New York GTM role) signals a research-and-product core in the U.K. and a commercial beachhead in the U.S., consistent with a team that went through Y Combinator Summer 2024.
For technical tracks (research engineer, full-stack product engineer, forward deployed engineer), role specs show the bar: production-grade LLM experience, comfort shipping to regulated environments, and the ability to own a slice of the "adaptive decisioning infrastructure" the company describes. A strong application shows shipped systems — not notebooks — and references to evaluation frameworks, guardrails, or compliance-adjacent work (financial crime, lending, KYC/AML). The Data Associate role requires labelling and evaluation discipline, signaling the research loop includes human-in-the-loop quality gates candidates should be ready to discuss.
Go-to-market roles (Enterprise Account Executive, Strategic Customer Success) ask for enterprise sales cycles and financial-services fluency. The account executive role implies quota-carrying experience selling to banks or fintechs; customer success looks for post-sale ownership in a high-trust, regulated setting.
No public benefits doc exists; ask the recruiter. The Greenhouse portal is the only official application route; third-party aggregators (GenZCareer) currently show zero indexed postings.
Two Offices, Two Regulatory Perimeters
Arva AI operates from two core offices (London and New York), placing the team inside the regulatory and customer ecosystems the company serves. The about page states the team works "from London and New York, close to the customers and regulators we build for," and board data confirms every current opening is based in one of those two cities. A third location, San Francisco, appears in some third-party directories as a headquarters, but live job postings and recent LinkedIn activity show no roles there; the operational footprint is London and New York.
London: Research Lab, Product, and European Regulatory Access
The London office occupies 93 Gloucester Place in Marylebone, a serviced space the company moved into as "a larger space built for where the team is headed next," according to a September 2026 LinkedIn update. That move coincided with the launch of Arva Research, the internal lab building proprietary models and the AgentCore infrastructure that turns analyst corrections into versioned system improvements before they reach production. The lab's first outputs are already running inside global financial institutions, including a top-10 US bank, the company announced September 10, 2026.
The concentration reflects the office's dual mandate: build the models that automate high-stakes compliance decisions, and ship product that integrates with the FCA, OCC, FDIC, and SEC frameworks the company certifies against (SOC 2 Type II, ISO 27001, ISO 42001). Third-party data notes London "provides access to the European financial services ecosystem and regulatory expertise." That proximity shows up in the customer roster: Blockchain.com, a long-standing crypto platform, moved decisions that previously required manual analyst review into governed automation using Arva; Wise, operating across dozens of jurisdictions, is onboarding with the platform. Both are European-headquartered or European-heavy institutions.
New York: Enterprise GTM, Fiserv Partnership, and US Bank Integration
New York carries the go-to-market weight. The Enterprise Account Executive role is based there, and company posts confirm the GTM team's physical presence: Cole Draghi joined the New York GTM team in October 2026 after five years at ADP, brought in to "guide every deployment from the outset to going live" with compliance and risk leaders at enterprise banks.
The partnership with Fiserv, announced June 2, 2026, runs through New York. Fiserv's network of banks and credit unions is a US-centric distribution channel; putting the enterprise team in New York shortens the sales cycle and the integration timeline. The Compliance Collective (an industry event series Arva convenes) reconvened in New York on October 15, 2026, "to celebrate the launch of Arva AI's Research Lab" with compliance and banking leaders in the room. That event signals the office's role as a convening point for the US regulatory and buyer community.
Hybrid Model and Founder Proximity
Both offices run hybrid setups. The about page lists "London & New York: Hybrid working setups supporting our core offices" and "Proximity: Work directly with founders, customers, and the compliance experts we build for." The small team means office days function as de facto sprint planning: researchers, product engineers, and forward-deployed engineers share physical space with the compliance practitioners who validate model outputs before they hit a bank's production queue.
The geographic split mirrors the regulatory split: London for FCA and European framework alignment, New York for OCC, FDIC, SEC, and the Fiserv partner channel. Candidates choosing between the two aren't just picking a city; they're picking which regulatory perimeter and customer segment they'll sit inside daily.
Who Thrives at Arva AI
The company's cultural statement makes the bias clear: "Deliver Value Fast. Speed starts with clarity. We first understand what value actually means, for the customer, the business, or the system, and then take the shortest credible path to delivering it." That phrasing — "shortest credible path" — reveals a bias toward engineers who distinguish motion from progress. It explains why the Forward Deployed Engineer role demands three-plus years in customer-facing software or solutions engineering, with explicit comfort across APIs, data pipelines, and modern web backends. The person who thrives here doesn't wait for a spec; they ship the integration that unblocks the customer, then iterate.
The hiring record shows ownership matters. When Tom Green joined as Senior Product Engineer, the announcement noted that "engineers who've built and run their own product bring a rare kind of ownership to a team." That pattern repeats across roles: engineers in London, sales in New York. The geographic split isn't arbitrary: it mirrors the product. Arva AI describes itself as "a research company deploying autonomous AI decisioning for critical banking operations, combining proprietary models, adaptive decisioning agents and controlled continuous-learning infrastructure." The London cluster builds the models and the platform; the New York cluster carries them into regulated financial institutions. People who thrive tend to have operated at that seam before: research that ships, not research that publishes.
The Head of Research and Data Associate roles bookend a team that takes evaluation seriously. "Controlled continuous-learning infrastructure" implies a feedback loop that runs in production, not offline. Candidates who have designed labeling workflows, built evaluation harnesses, or monitored model drift in regulated environments map directly to that need. The Strategic Customer Success role further signals that post-deployment iteration is a first-class function, not an afterthought.
Collaboration is explicit in the forward-deployed brief: "Partner closely with product, engineering, and leadership to align customer outcomes with long-term platform strategy." That triple alignment (customer, platform, leadership) filters for people who can translate a bank's compliance constraint into a platform primitive without losing either thread. Founders Shah and Wales set the tone; the team they've assembled across eight salaried roles reflects a consistent bar: strong fundamentals, production experience, and a default to the shortest credible path.
As Arva AI scales, the unusual intersection that defines it — research lab inside a commercial company, teaching agents to make decisions banks trust only to humans — only sharpens. The lab builds what the bank buys. The hiring bar ensures both sides speak the same language.
Working in AI? Zero G Talent tracks the openings: see every open Arva AI role, browse AI jobs, the companies hiring, and the people building the field.