Full Stack Engineer

Senior Full-Stack Engineer (AI-Native)

• Remote from Vietnam
• Australian SaaS company
• Contract Role, with strong potential to go long-term

We are on the hunt for a Senior Full-Stack Engineer, the AI-native kind.

If your working day still looks much like it did before the models arrived, you code by hand with Copilot on, you ship large generated changes, you work to scoped tickets, or you lean on the model as your verification, this role isn't for you.

In this role you own the product outcome, not a task list.

The platform models complex financial outcomes and its users act on those numbers commercially, so accuracy is not a quality bar, it is the product.

The job is deciding what should exist, directing the machine that builds it, and proving the result is right.

What AI-native means here

• Your output has multiplied, and coding has been replaced with deciding and verifying
• Your workflow is built around the models, not interrupted by them
• Your patterns, context and instructions get models producing correct work first time
• You write for a model with precision, because ambiguity is where it goes wrong
• You know where a model quietly does the plausible thing instead of the correct one, and you check there first
• A model can generate a calculation function. Proving it is right is the part of your role

The role

Reports to the Lead Engineer and the CTO, working directly with the founders and product team.

A product role as much as an engineering one. You own features end to end. Intent comes from a problem to solve, not a ticket, and often the work starts with you: a customer question, a support pattern, an output that looks off, or a part of the platform you think can be better.

Shipped is not finished. Your instinct on a user question or a piece of feedback is to look for the better solution, even when the work went out two months ago.

Curiosity matters as much as capability: how the product is used and why the numbers matter to the people using them, not only how the code works.

What you will be doing

• Turning customer questions, feedback and platform behaviour into defined problems, then assessing the options, ruling things out and putting up a clear proposal
• Investigating the existing system before changing it, tracing an input through the full model rather than fixing where the symptom surfaced
• Planning and breaking work into small, reviewable, reversible changes. Planning is what makes speed safe, not a substitute for it
• Directing AI implementation across the frontend, backend and shared calculation engine, reading every generated change before it ships
• Finding the edge cases and risks early, and raising them in writing
• Verifying outputs with evidence, independent of the model that produced them
• Reviewing your own work critically before it reaches users, and owning the result after release
• Proposing improvements to work already in production

How this team works

• Delete before adding. Abstraction only when something real calls for it
• Arguments come from evidence. Not "the model suggested it"
• Ask early and loudly, but after tracing the code, not instead of it
• Don't bypass dull procedures, suggest to change them.
• Staying quite doesn't work. Made a mistake, need help ask the team.

Core stack

Business logic and calculations live server-side. The frontend collects inputs and presents results. You need to be strong across:

• TypeScript
• React and Vite
• Node.js
• PostgreSQL

Is this you?

• AI-native working as described above. This is non-negotiable.
• 5+ years of professional software engineering
• Strong full-stack capability across modern React and Node.js
• Solid TypeScript and relational database knowledge
• Experience with complex, interconnected business logic
• Genuine testing and verification discipline
• Professional written and spoken English, non-negotiable

Next step

Apply for the role. Shortlisted applicants will be sent a short set of questions about how they work, how they verify, and how they use AI. They are straightforward to answer well if this is genuinely how you operate, and difficult to fake if it is not.