Nylon is the AI research platform for tax and law professionals.
The tool practitioners reach for when the answer has to be right. We're used by 2,000+ firms across NZ, AU and the UK, backed by Blackbird, and expanding fast beyond research into the wider workflow professionals live in: client meetings, documents, the tools firms already use.
The roleWe’re hiring senior full-stack engineers (backend-heavy) to own product surfaces end to end: architecture, delivery and operational excellence. You’ll work with product from “should this exist?” through to running in production, then keep iterating on real usage. Your work lands with thousands of firms across NZ, AU and the UK.
What you could ownWe're hiring for four key surfaces. You'll have one as your primary (we'll agree which during the process) and work across the others.
The answer experience: the heart of Nylon. How a professional asks a hard question and trusts what comes back. Retrieval, citations, verification, speed.
Identity, billing & the public API: make Nylon work as well for a 500-seat firm as for a sole practitioner. SSO, entitlements, usage metering and the API other platforms build on.
Integrations: the wedge into the daily workflow. Outlook, Xero, practice-management systems. Each one pulls Nylon deeper into how a firm actually runs.
Workflows: help professionals engage clients and gather the facts behind the advice. Meeting AI is the first: already in real client meetings, capturing conversations and connecting them back to research.
AI-native. Agentic dev tooling (Claude Code) is the development loop.
You run what you ship. Owning a surface includes its production health, observability, performance, incident response.
Async-first. Clear writing is a first-class engineering skill here.
No prescribed solutions. We agree the problem and the constraints; how you solve it is yours.
We care about evidence of ownership, though you've likely got 5+ years of commercial software development experience.
You can point at a system you took from design to production and then operated
High agency: biased toward action, you own outcomes, not tasks
You decompose hard problems and reason in trade-offs, out loud and in writing
You already use AI tooling to ship faster, and you're curious about where that goes next
You build robust systems to reduce the 2am pagers
Node.js
PostgreSQL
AWS
Terraform
Jest
Docker
Datadog
GitHub Actions CI/CD
React
.png)
