AI Engineer / Full-Stack Builder - Industry Apps

Most AI roles are narrow: a model, a pipeline, a ticket. The interesting problems in the resources industry aren't narrow at all. They're tangled, half-defined, and sit across the full spread of an operation's data, from clean telemetry to the messy signal that never made it into a system. The work is taking that operational data, applying fast-moving tools in ways they haven't been applied before, and turning it into something that delivers real value. That takes invention, not just implementation. This role is for the builder who'd rather find a way through a problem nobody has solved than supervise autocomplete in a boardroom.

Industry Apps makes software for mining and resources. We have products deployed and running at Tier-1 sites today, and we're building the next generation of AI-driven tools on top of that foundation. You'd help take real operational problems from idea to working, deployed systems, treating AI as a sharp tool to be aimed carefully rather than a hammer for every nail.

Who you'd work with

You'd build directly alongside a founder who understands the software, the AI, and the customer's actual problems, so there's no translation layer and no one between you and the decision. We move fast, test ideas in front of real users, and back the ones that work. The aim is novel, high-value products grounded in genuine customer pain, not features for their own sake, and you'd have a real hand in shaping what we build and how.

Why this is different

A lot of the work is figuring out what to build, not just how. The problems are real and the customers are real, but the best solutions usually come from an unexpected angle. If you think laterally and you're at your best when the path isn't fully drawn, you'll have room to run, with the context and support to make good calls rather than guess in the dark.

You probably

• Built and shipped real software the conventional way before AI was the headline, and carry the production scars to prove it: data integrity, edge cases, and the long gap between a slick demo and something that holds up in the field.

• Moved into applied AI and LLMs since, and learned to trust a model as a component while staying clear-eyed about where it breaks and when to leave it out entirely.

• Are comfortable working across the stack and enjoy carrying a problem from idea through to something real, rather than owning only a slice of it.

• Reach for unusual solutions, and have a habit of finishing things.

What you'd actually do

• Take real operational problems and help shape and build the systems that solve them, often where there's no template to copy.

• Decide where AI genuinely earns its place and where simple, boring code is the smarter call.

• Build across the stack and care that what ships is correct, observable, and honest, because the output drives real decisions on real sites.

• Get something working quickly, put it in front of people, and improve it from contact with reality rather than from a spec.

What we care about

• Proof you've built things that shipped and worked.

• Creativity and lateral thinking applied to hard, underspecified problems, backed by the engineering to deliver.

• Real fluency with modern AI, paired with healthy skepticism. You know its failure modes, not just its highlight reel.

• Sound judgment, because the team is small and the work moves quickly.

The toolkit (we care more about what you've made than which you've used): LLMs and agentic approaches, a modern full-stack web setup, and the cloud or on-device platforms real operations rely on.

The setup

• Remote within Australia, with the door wide open to someone based in (or keen to be in) Port Hedland or the Pilbara. We're on the ground here ourselves, so if you want to be close to the operations, the customers, and the problems, all the better. If remote suits you better, that works too.

• $180–230k+ depending on experience, with equity for the right person.

• Products already in the field, customers at Tier-1 sites, and a genuinely short road from idea to deployment.

If the thought of building things that act in the real world beats supervising autocomplete in a boardroom, send us what you've built.