Forward Deployed Engineer
My client is a consulting firm building AI systems that go live inside their customers' businesses — not slideware, not pilots that quietly stall after the workshop.This is a Forward Deployed Engineer role, which means you sit with the customer. You'll be in their offices, in their data, in the meetings where the actual problem finally surfaces — usually about three weeks after the brief said something else entirely. Then you build the thing, ship it into production, and stay long enough to know whether it worked.
This role suits an engineer who is genuinely good with people and slightly allergic to abstraction. You'll write real code across the full stack, make architecture calls without a committee, and carry the consequences of both. Every few months, a new customer, a new domain, a new mess to make sense of.
This is a 12-month contract engagement based in Melbourne, at $1,250 per day.
What you'll do
- Embed with customer teams to work out what they actually need, then design and build it end to end
- Ship production applications across the full stack — front end, services, data, deployment
- Integrate LLMs, retrieval, embeddings and agent frameworks into systems people rely on daily
- Stand up secure, scalable cloud architecture and own it through to monitoring and iteration
- Build the operational layer around models: prompt and version management, pipelines, evaluation, cost and latency tuning
- Work with messy real-world data — text, documents, images, audio — including everything nobody warned you about
- Translate between the customer's language and the engineering team's, in both directions
- Lift the standard around you: reusable components, sensible patterns, mentoring, honest code review
What you'll bring
- Strong full stack engineering, with real production depth on both sides of the stack
- Hands-on experience taking GenAI or LLM-based features into production, not just prototypes
- Solid cloud skills and the judgement to know when a simpler architecture is the right answer
- Confidence in front of customers — including when the news isn't good
- Comfort with ambiguity, shifting scope and the occasional half-defined brief
- The instinct to ask what problem this actually solves before writing a line of code
Nice to have: consulting or client-facing delivery background, exposure to regulated environments, and experience mentoring engineers.
