AI Research Engineer

AI Research Engineer

Do research that ships, and build the systems around it.

arbitr isn't asking you to fine-tune someone else's roadmap. We're asking you to help define it.

We build models that run inside our platforms. We build the retrieval, memory, and evaluation systems around them, and we ship the whole thing into production, where it has to actually work.

arbitr is a 25-year global company (formerly Straker Ltd) — ASX-listed, offices across the USA, Japan, Europe and New Zealand - that has always been at the forefront of technology, and that includes AI, from the inside out. We operate in smaller teams, as a startup, which means you get something rare: the scale and stability of an established business, with the mandate and the appetite for risk of a team that is genuinely starting again.

If you want to do research that ships, not research that lives in a notebook, this is that role.

The team

You'll join our AI research and engineering team, reporting to and collaborating with the Head of AI Research and Model Engineering. We build the models that power arbitr's platforms, we build the systems that give those models the right context and structure to be useful, and we act as the AI experts across the company and for our clients, providing technical guidance to product, engineering, and partnerships.

What you'll own

• Build custom models. For customers, and across arbitr's platforms. You own the pipeline end-to-end - first experiments through production and beyond.
• Build the systems around the models. Retrieval. Memory. Reasoning. Orchestration. Whatever the model needs on either side of it to actually deliver value in production.
• Own our evals. Build the benchmark tooling that tells us whether a model is genuinely good, not just good on paper.
• Push the research. Track new models and training techniques as the field evolves. Bring the ones that matter into our stack.
• Design the research workstreams. Take an idea from a problem statement, through experiments, findings, and the pipeline changes on the other side.
• Deploy and run models at scale. Serve, monitor, and orchestrate models on our GPU infrastructure. Production is your problem to own.
• Work at the partner table. Support strategic engagements with global partners, where your technical judgement carries weight.
• Go wide across arbitr. Bring engineering support to arbitr's wider platform work: architecture reviews, AI strategy input and hands-on development where it counts.
  

You'll thrive here if

You've trained a model before. Not fine-tuning someone else's on HuggingFace, but actually training from the ground up. You understand what's happening under the hood and you can debug when it goes wrong.

You think in experiments, not vibes. You build real evaluation gates before you trust a result.

You are relentless about getting from idea to production. You filter hard for what is signal versus noise. You want to join a team in the middle of reinventing itself, and leave your work in the shape of what comes next.

What we're looking for

• Detailed understanding of model building - concepts like LoRA, PT, reinforcement learning, and inference are familiar territory.
• Comfort designing and building datasets, and a working background in data science.
• Ability to work alone on a project, or with a team of like minded individuals.
• Strong Python skills.
• A self-starter mindset. You are part of a team, but personally accountable for outcomes.
• Familiarity with the ideas we write about at, or the appetite to get familiar quickly.

Practical details

• Location:Auckland, NZ / hybrid
• Work rights: Must have working rights for NZ
• Compensation: competitive base + benefits
• Diversity: we hire for judgement, curiosity and evidence, not for pedigree. We actively welcome applicants from backgrounds under-represented in AI research.
• Accessibility: if you need accommodations at any stage of the interview process, tell us and we'll make them.

Why arbitr, why now

25 years of proven results. A global footprint across the USA, Japan, Europe and New Zealand. And a company that has just renamed itself around the work it now does - with a new mandate to push to be AI-native from the ground up.

We're building the infrastructure, research systems and shared layers to move with a field that changes month to month. If you want to help shape what that foundation looks like, we'd like to hear from you.