Senior Software Engineer
Salary: $120,000 – $160,000 per year
We build software with agents, safe and fast.
Most engineering teams are experimenting with AI coding tools. We've rebuilt how we work around them.
On any given week, most of the code that lands in our main repo was written by an agent. Our engineers spend their time on the parts that actually decide whether the product is good: what to build, how it should be shaped, whether the thing that came back is correct, and what guardrails stop the next one from getting it wrong. We ship dozens of pull requests a week into a merge queue with automated gates on architecture, security boundaries, test quality, and file-level complexity - because at this throughput, review discipline can't be a matter of remembering.
This is an in-production, regulated financial platform, built by a team of 10 and growing.
What Adiso is
Adiso is the operating system for New Zealand's financial advice industry - CRM, compliance, workflow automation and advice modelling in one platform. KiwiSaver, mortgage, insurance, investment. Our customers are wealth networks, boutique advice firms, KiwiSaver providers and independent advisers who currently lose enormous amounts of time to compliance paperwork and disconnected systems.
The domain is genuinely hard and worth getting right. Advice records are legal artifacts. Statements of Advice have to be reproducible years later. Calculations inform real investment decisions. Personal data cannot leak. Security work is a first-class part of the roadmap rather than a compliance afterthought.
The stack
TypeScript end to end: React on the front; Node/Express and PostgreSQL with Drizzle on the back. Vitest and Playwright. GitHub Actions, Sentry.
We build with Claude and Codex.
What you will be building
Adviser based financial advice software for various domains - advisers work with this software all day, every day
An in-product AI assistant with tool access to client data, meeting transcription & summaries
Mobile applications for end clients receiving financial advice
Integrations with industry partners
Safety and integrity checks
What we're looking for
You've shipped production software that other people depend on, and you can own a feature from schema to UI
You can design a feature up front, with strong architectural consideration, solving a problem once for the codebase, not per-file quick fixes
You have strong opinions about correctness - tests that assert behaviour rather than decorate coverage, failure modes that fail loudly rather than guess a plausible default
You can review code critically and quickly, and write unaided anything you'd accept from an agent. Review is a large part of this job, not an overhead on it
You're already using AI coding tools seriously, and you have a real sense of where they're strong and where they'll confidently hand you something wrong
You can read a problem in an unfamiliar business domain and ask the questions that reveal the hidden requirements
You write and communicate clearly.
We haven't put a years-of-experience bar on this. If you're early in your career but can demonstrably do the things above, apply.
What you get
Real AI budget. We spend heavily on frontier model access for our engineers and we don't ration it. No expensing your own subscription, no queuing for a seat, no asking permission to use the good model on a hard problem.
Authority over how we work. Our workflow is still changing month to month. If you can show us a better gate, a better review shape, or a better way to drive an agent, you change it - it’s a core part of the job.
Skills that compound in the new paradigm - architecture, correctness, judgement - rather than volume of code written.
How we work
Small branches, behind feature flags, reviewed within hours, merged within a day. Architectural decisions get written down as ADRs. Every PR carries an explicit risk assessment. When a rule matters, we make it a gate rather than a wiki page nobody reads.
You'd be joining a team that includes people with deep advice-industry experience alongside the engineering side, so the feedback loop between "what advisers actually need" and "what we built" is short.
To apply
Send us your CV and a short note on something you've built or fixed that you're proud of, what was hard about it, and how you knew it was right. If you've been using agents in your own work, tell us what your setup looks like.
Include your salary expectations, location, work rights.
We hire on how you think and what you can show us. If this sounds like the way you already want to work, we'd like to hear from you.