RAG is easy to demo. Hire the engineer who has made it right on the hundredth question.
FERG Search is a specialist recruiter for Azure AI engineers in Australia and New Zealand, permanent and contract: pro-code engineers on Microsoft Foundry (formerly Azure AI Foundry) and Azure OpenAI who build retrieval augmented generation, AI search and agents for Microsoft partners and end users. Every search is run by the founder, Fintan Ferguson.
A retrieval pilot answers the first ten questions well. The trouble starts with the next ninety: stale documents, permissions the index ignored, answers nobody can trace back to a source, and a token bill nobody forecast. The engineers we place have been through that stretch and can show what they changed.
First shortlist within 48 hours of an agreed brief.
Engineers who have shipped AI on Azure.
- Senior AI engineers on Microsoft Foundry and Azure OpenAI.
- RAG and AI search engineers: chunking, embeddings, Azure AI Search and evaluation.
- Agent engineers who build and orchestrate agents in Microsoft Foundry.
- AI engineers who connect Foundry agents to Copilot Studio, Dataverse and Microsoft 365.
- Python and .NET engineers moving from application work into AI.
Checked technically and commercially, before you meet them.
- The retrieval system they shipped, and how they measured whether the answers were right.
- How the index respects the permissions on the documents underneath it.
- The token and capacity cost, and what they did when it ran over.
- The evaluation and monitoring that runs after go-live, and who reads it.
What the seat pays.
AI engineering briefs stall when a Foundry seat is budgeted at a Copilot Studio band.
Senior AI engineer, pro-code on Azure and Microsoft Foundry: $170k to $210k. Sydney. Melbourne $165k to $200k. Contract $900 to $1,200 a day.
Copilot Studio developer and agent builder: $140k to $175k. Sydney. Contract $800 to $1,050 a day.
The step from a Copilot Studio developer to a pro-code AI engineer on Microsoft Foundry is $30k to $35k on the bands above. Permanent figures are base salary excluding super. Contract figures are per day excluding GST. Sydney unless stated; Melbourne generally runs 4% to 6% lower, and where a Melbourne band is shown, that band applies.
The platform changed its name. The skills did not.
Azure AI Foundry is now Microsoft Foundry, and Azure OpenAI models sit inside it. Job descriptions in Australia and New Zealand still use all three names, often in the same advertisement. We search on what the engineer has built, not on which name was on the portal when they built it.
Straight answers on Azure AI engineers.
Who recruits Azure AI and Microsoft Foundry engineers in Australia and New Zealand?
FERG Search recruits Azure AI engineers, Microsoft Foundry engineers and RAG engineers for Microsoft partners and end users across Australia and New Zealand, permanent and contract. The search is run by the founder, Fintan Ferguson, from Melbourne.
What does a senior AI engineer earn in Sydney or Melbourne?
The FERG Search Salary Guide 2026 puts a pro-code senior AI engineer on Azure and Microsoft Foundry at $170k to $210k base in Sydney and $165k to $200k in Melbourne, excluding super, or $900 to $1,200 a day on contract excluding GST.
Is Azure AI Foundry the same as Microsoft Foundry?
Yes. Microsoft renamed Azure AI Foundry to Microsoft Foundry, and Azure OpenAI models are offered inside it. An engineer with Azure AI Foundry or Azure OpenAI on their CV has worked on the same platform.
What should a RAG engineer be able to show?
A retrieval system in front of real users, the way they measured answer quality, how the index handles document permissions, and what it costs to run each month. A demo on a clean document set shows none of those.
Do I need an AI engineer or a data engineer first?
If the documents and data the agent will read are not in order, the data engineer comes first. An AI engineer can build retrieval over messy content, but they will spend the first months cleaning it, at an AI engineer’s rate.
Brief an AI engineering role.
Tell us what the system has to answer, from which data, and who checks that it is right.