AI / ML Engineer

Production AI engineers who ship. Sonitec recruits the AI and ML engineers who take models into systems real users rely on — across Australia and APAC.

The seat, defined

The AI / ML Engineer is the seat that turns a model into a product. The title varies — ML Engineer, AI Engineer, Applied Scientist with a delivery remit — but the constant is production: systems that real users rely on, with the reliability, cost and latency characteristics that implies. Whether the work is classical ML, fine-tuned models, or LLM-backed applications, the job is the same shape — build it, prove it, run it.

It is a different seat from research. A researcher is measured on what is novel; an AI / ML Engineer is measured on what is shipped. The strongest people in this seat treat a notebook as a starting point, not a deliverable.

Shipped is the bar

The gap between a working prototype and a production AI system is where most AI initiatives stall — and it is exactly the gap this seat exists to close. That means:

Eval-driven development — regression suites and benchmarks that prove behaviour before and after every change, not vibes-based sign-off.
Operating in production — monitoring, failure handling, cost and latency budgets, and ownership of the system after launch.
Grounding with data — retrieval and pipelines built against the organisation’s real data, with its real quality problems.

Coding agents are part of how this seat works now — a strong AI / ML engineer directs them through the build and stays accountable for what ships. The tooling is the method, not the title.

Seat Context

The skills behind the seat

We frame this seat against Andrew Ng’s AI Engineering Skills Map — four parent skills, weighted here toward the build-and-deploy core. Context for the seat, not a checklist.

Building & deploying AI applications

The core of the seat — applied, production-facing:

ML foundations — training, evaluation, and knowing which technique fits
LLM foundations — working knowledge, applied not academic
Grounding with data — RAG, retrieval, and data pipelines
Agentic systems — tool calling, memory, structured outputs
Eval-driven development — behaviour proven, not asserted
Operating in production — monitoring, guardrails, cost and latency

Software engineering fundamentals

Python and/or TypeScript, APIs, cloud, testing, version control. Production AI is software first — this is what separates an engineer from a notebook.

Using coding agents effectively

Working through real codebases with agents at production pace — while owning the review, the evals, and the outcome.

Shaping the build

Working with product owners and domain experts on what to build and where AI earns its place. Less central than in the FDE seat, but the best engineers in this seat do it anyway.

How We Qualify

What we interview against

Sonitec is a specialist AI recruitment agency. We staff this seat — we don’t build your AI. Every AI / ML engineer we put forward is interviewed against the seat as defined on this page.

Evidence of shipping

Real AI systems used by real users — with the candidate able to walk through what broke, what it cost, and what they changed.

Eval discipline

How they proved the system worked — the datasets, the regressions, the benchmarks — and how they caught it when it stopped working.

Engineering depth

The software fundamentals under the AI work — because a model that can’t be deployed, monitored, or maintained isn’t a product.

Hiring around this seat too? See Model Development and MLOps & AI Infrastructure for the specialists we place alongside AI / ML engineers, and the salary guide for current AUD benchmarks.