Senior AI Engineer — Agentic AI (Azure AI Foundry / Databricks)
Data Diversity Pty Ltd · Sydney NSW ·
- Work mode
- Hybrid
- Seniority
- Senior
- Employment
- Full time
- Category
- AI engineering
- Experience
- 5+ years
Data Diversity Pty Ltd · Sydney NSW ·
Salary: $110 – $125 per hour
Client: Leading Australian insurance group (name disclosed at shortlist stage)
Location: Sydney, Australia (hybrid - client site + remote)
Employment type: Contractual Experience: 5+ years
Eligibility: Australian citizens and permanent residents only
We are recruiting on behalf of a leading Australian insurance client who is scaling agentic AI from proof-of-concept into production across their business. This is a hands-on engineering role with a delivery leadership edge - you will design the solution, build the core of it, and then stand in front of business stakeholders and explain why it works.
We are looking for someone who has already delivered this at least twice end-to-end, and who understands what it takes to ship AI inside a regulated financial services environment where architecture review, risk sign-off and audit trails are part of the job, not an afterthought.
Design and build agentic AI solutions using LangGraph (or equivalent orchestration frameworks) on Azure AI Foundry and/or Databricks.
Own the technical solution end-to-end — architecture, data pipelines, agent orchestration, tool/API integration, evaluation, observability and production hardening.
Apply agentic patterns to insurance workflows: claims triage and assessment, underwriting support, policy and document intelligence, customer correspondence, broker servicing and regulatory reporting.
Lead delivery of AI workstreams: shape the approach, sequence the work, unblock the team, and hold the quality bar.
Run working sessions with the client's business and technical teams — translate ambiguous business problems into implementable designs, and translate technical trade-offs back into business language.
Produce solution design documentation and architecture artefacts that pass the client's enterprise architecture, security and risk review gates.
Present solutions, demos and outcomes back to business sponsors and executive stakeholders.
Embed responsible AI practice — evaluation frameworks, guardrails, human-in-the-loop patterns, explainability, model governance and cost/performance monitoring.
Minimum 5 years in data analytics, big data, AI/ML, Databricks or Azure cloud engineering.
Minimum 2 hands-on projects delivered end-to-end where you personally built and shipped the solution — not advisory or oversight-only involvement.
Demonstrable production experience with Azure AI Foundry and/or Databricks (Unity Catalog, Delta, MLflow, workflows).
Practical experience building agentic / multi-agent systems with LangGraph or a comparable framework — state management, tool calling, retries, memory, evaluation and failure handling.
Strong Python engineering fundamentals, plus RAG patterns, vector stores, prompt engineering and LLM evaluation.
Proven understanding of what "enterprise" actually means: identity and access, data residency, network boundaries, CI/CD, change control, cost governance and support models.
Confidence operating inside a large client organisation — navigating their governance, working alongside their engineers, and earning trust with their business stakeholders.
Ability to lead solution conversations independently with both business and technical audiences, and to present with credibility.
Unrestricted working rights in Australia — Australian citizenship or permanent residency required, as the engagement is subject to the client's background and financial services vetting requirements.
Insurance domain experience — general, life or health insurance; exposure to claims, underwriting, policy administration, broker channels or regulatory reporting.
Familiarity with the Australian regulatory context: APRA CPS 230 and CPS 234, Privacy Act obligations, and internal model risk or AI governance frameworks.
Azure Functions, API design and integration (REST, event-driven), Kafka / Event Hubs and streaming architectures.
Azure DevOps, Terraform/Bicep, containerisation.
Relevant certifications (Azure AI Engineer, Databricks Data Engineer/ML Associate or Professional).
In the first six months you will have taken at least one agentic solution from discovery through to production in the client's environment, built their confidence in how it behaves, and left behind design documentation and an evaluation harness their team can run without you.