Lead Applied AI engineer

About Fetch

Fetch is how modern pet parents give their pets the best care. One pink app for insurance, health and care.

We've raised our Series A from Lightspeed and Airtree, we're rated 4.9 ️ on Google, by CHOICE, and we're growing fast.

We think AI is going to fundamentally change how we care for pets.

We have pet health data, an in-house team of vets and vet nurses, and we build our own product and technology. We want to turn that into an intelligence layer that understands each pet deeply - their history, health and what might matter next.

We're looking for an exceptional Lead Applied AI Engineer to build it.

The role

This is a senior, hands-on builder role.

You'll lead applied AI at Fetch - setting the direction, choosing the models and approaches, building them yourself, measuring whether they actually work, and shipping them into real products.

Your core job is simple: make Fetch exceptionally good at understanding pet health.

That capability will power everything from helping pet parents care for their pets, to supporting vets, interpreting health records and automating parts of insurance.

We don't need someone with a particular academic pedigree. We want someone who is obsessed with what modern models can do, moves quickly, experiments constantly and knows how to turn new capabilities into reliable products.

• Location: Sydney, hybrid, 3 days in office

• Employment type: Full time

• Compensation: Series A salary + meaningful equity

• Working with: Co-founders, engineering, product, data, vets and vet nurses. We're flat, fast and care about the best idea.

What you'll build and own

• Fetch's pet health intelligence layer: models and systems that can understand messy, longitudinal pet health information and reason over it

• Model quality end-to-end: datasets, evals, model selection, prompting, retrieval, post-training, fine-tuning, guardrails and monitoring

• A serious evaluation system: clear benchmarks for what good looks like, where models fail and whether a new model is actually better

• Domain-specific AI: turning proprietary pet health data and expert veterinary knowledge into capabilities generic models don't have

• Multimodal understanding: health records, documents, images and, over time, richer inputs like video

• Safe production AI: working closely with our vets, compliance and engineering teams to make sure the systems we ship are reliable and clinically appropriate

• The next generation of Fetch products: you'll be deeply involved in product direction and help decide what we should build with AI, not simply implement someone else's roadmap

• The beginnings of an exceptional AI team: this is a core capability for Fetch and we expect the team around it to grow

You'll use whatever approach works best.

Sometimes that will mean a frontier closed-source model. Sometimes an open-source model. Sometimes RAG, tools or better context. Sometimes fine-tuning or post-training.

We care about getting to the best result, measuring it properly and improving it quickly.

About you

You're an exceptional applied AI builder who wants a lot of ownership.

You've probably spent the last few years building with modern ML and LLMs and have gone much deeper than plugging an API into a product.

This might look like:

• 3+ years building and shipping ML or AI systems in production

• Strong experience with modern LLMs and applied AI systems

• Deep understanding of evals - how to construct them, measure model quality and use failures to drive improvement

• Experience with post-training, fine-tuning, RAG, structured outputs, tool use or similar techniques

• Strong instincts around training data - what data matters, how to structure it and how to improve its quality

• Enough engineering ability to take an idea from experiment to production

• Good judgement about when to use a frontier model, an open-source model, retrieval, fine-tuning or something simpler

• Strong product thinking - you care what the system enables for the user, not just the benchmark score

• Comfortable working directly with domain experts to turn human expertise into data, evals and model behaviour

Experience in health, veterinary care or insurance is us