The deployment gap
Most enterprise AI programmes don't fail because the model is bad. They fail because nobody owned the distance between a general-purpose AI capability and the messy, exception-riddled reality of a specific business. The model works in the demo; it dies in the workflow. The role that closes that gap has a name — and it has quietly become the fastest-growing job family in applied AI.
What a Forward Deployed Engineer actually does
A Forward Deployed Engineer (FDE) is an engineer a company embeds directly with a customer — often on site — to turn general AI capability into working software for that customer's exact business context. The title was popularised by Palantir; today OpenAI, Anthropic, Google, Salesforce, Databricks and the major consultancies all hire under it.
The work runs in three stages. First, understand the business reality: sit with the people doing the work and map the real workflows — not the documented ones — including all the exceptions living in people's heads. Second, apply deployment judgment: decide which steps genuinely need an LLM, which should stay deterministic software, and where a human must remain in the loop. Third, build, evaluate, deploy: ship production-grade agents with evaluation suites, guardrails, audit trails and monitoring, integrated into the systems the business already runs on — ERP, CRM — rather than replacing them.
Why the role is rare
The FDE profile demands the best of two skill sets that seldom coexist: consulting-grade communication and stakeholder management, plus production-grade AI engineering. A brilliant researcher may never have sat with a warehouse team; a brilliant consultant may never have shipped an agent past a proof-of-concept. The people who can do both are scarce everywhere — and top FDEs command extreme compensation as a result.
The Australia & New Zealand picture
Australia trails the US on the FDE hiring wave. As of August 2026, independent public counts put the number of distinct roles advertised under the exact FDE title in Australia at around a dozen — Databricks, Anthropic, Salesforce, Google, Deloitte, Cloudera and Accenture among them — and New Zealand has virtually none. Yet. The searches that do run close fast: published Q2 2026 market data has the median FDE search at 21 days from brief to signed offer, the fastest at 14. Against that sits Australia's projected shortfall of roughly 60,000 AI professionals by 2027, per Bain & Company research — and the FDE is the sharpest end of that shortage.
Because the title is new here, most FDE-capable people don't call themselves FDEs — knowing where they actually work is the sourcing edge. Supply has to be built, not found, and that is exactly what Sonitec is doing with ANZ's first dedicated FDE talent pool.
What FDEs earn
In Australia in 2026, published market data puts senior FDEs at $180k–$240k base in Sydney and lead or principal FDEs at $230k–$280k (AUD base, excluding 12% super); Melbourne runs slightly below and Brisbane lower again. The discipline is young enough that no junior or mid-level bands are published yet. Financial services and frontier AI scale-ups pay top-of-band, and production LLM deployment experience is the strongest premium driver. Full breakdowns are in our FDE salary benchmarks.
What to do with this
If you're a leader whose AI pilot is stuck between demo and production, the missing ingredient is probably not a better model — it's an FDE. If you're an engineer who has shipped LLM systems to real users and can hold a room of executives, you may already be one, whatever your title says. Either way, we should talk.
Common questions
What is a Forward Deployed Engineer?
A Forward Deployed Engineer (FDE) is an engineer a company embeds directly with a customer — often on site — to turn general AI capability into working software for that customer’s exact business context. The title was popularised by Palantir; today OpenAI, Anthropic, Google, Salesforce, Databricks and the major consultancies all hire under it. The work runs in three stages: understand the business reality by mapping the real workflows rather than the documented ones; apply deployment judgment about which steps genuinely need an LLM, which stay deterministic software, and where a human must remain in the loop; then build, evaluate and deploy production-grade agents with evaluation suites, guardrails, audit trails and monitoring, integrated into the ERP and CRM systems the business already runs on.
Why are Forward Deployed Engineers so rare?
The FDE profile demands the best of two skill sets that seldom coexist: consulting-grade communication and stakeholder management, plus production-grade AI engineering. A brilliant researcher may never have sat with a warehouse team; a brilliant consultant may never have shipped an agent past a proof-of-concept. The people who can do both are scarce everywhere, and top FDEs command extreme compensation as a result.
How many Forward Deployed Engineer roles are there in Australia and New Zealand?
Australia trails the US on the FDE hiring wave. As of August 2026, independent public counts put the number of distinct roles advertised under the exact FDE title in Australia at around a dozen, and New Zealand has virtually none yet. The searches that do run close fast — published Q2 2026 market data has the median FDE search at 21 days from brief to signed offer — against Australia’s projected shortfall of roughly 60,000 AI professionals by 2027 (Bain & Company).
Want to Discuss This Topic?
Whether you're hiring a Forward Deployed Engineer or want to join ANZ's first dedicated FDE talent pool, start here.