AI Engineer

St Vincent's Hospital Melbourne; Melbourne; Hybrid; AUD 120,000 - 200,000; Full time

Benefits offered by this job

Hybrid work (Melbourne)

Salary packaging options

Discounted private health insurance

Employee Assistance Program

Fitness Passport access

Job summary

St Vincent's Health Australia is seeking an experienced AI/ML Software Engineer to join the AI Delivery team in Melbourne. You will take AI solutions from prototype to production, integrating Databricks, Azure OpenAI and MCP in clinical and operational workflows.

You will design agentic AI, build scalable services, collaborate with governance and privacy teams, and contribute to the enterprise data platform while working in a hybrid model based in Melbourne.

Qualifications

• 5+ years in software, data or ML engineering.
• Ownership of at least one AI-enabled solution from prototype to production.
• Advanced Python engineering and one of TypeScript/JavaScript, C# or Java.
• Experience integrating large language model APIs into applications with authentication and error handling.
• Hands-on experience building agentic AI solutions with MCP and orchestration frameworks.
• Experience with vector databases or embedding models.

Responsibilities

• Build, test and deploy production AI solutions, turning architecture designs into maintainable code.
• Design and build agentic AI solutions using orchestration frameworks and MCP.
• Integrate LLMs via Databricks Mosaic AI Gateway and Azure OpenAI into workflows.
• Build on the CDAI platform, consuming governed data products from the Databricks lakehouse.
• Prepare structured and unstructured data for model development, evaluation and retrieval.
• Build and validate traditional ML models where suitable.
• Implement LLMOps and MLOps across the full lifecycle, including versioning and observability.
• Develop prototypes via the AI Incubator and contribute reusable components.
• Document patterns and mentor colleagues to build internal capability.

Skills

Advanced Python

TypeScript/JavaScript

C# or Java

APIs & microservices

SQL & data wrangling

LLM APIs integration

MLOps / LLMOps

Agentic AI / MCP

Tools

Databricks

Azure

MLflow

GitHub

Job description

Job Description

About the role

St Vincent's Health Australia has committed, through Strategy 2030, to connecting care across its public hospitals, private hospitals, aged care and virtual care services, with half of all care delivered virtually or in people's homes by 2030.

This role puts your Generative AI, software engineering and machine learning skills to work in healthcare. You will join the AI Delivery team within Data, Analytics, AI & Interoperability, report to the Senior AI Architect & Developer, and take AI solutions from prototype through to governed production.

You will:

• Build AI that clinicians, carers and staff use every day, and see its effect on patients and residents
• Join a new enterprise data and AI platform at the start and shape the patterns others will reuse
• Work with a current stack, including Databricks, Microsoft Azure, MCP, Generative AI Models, and GitHub
• Deliver alongside clinical, governance, privacy and legal colleagues, so your solutions are built to reach production

Key responsibilities of the role

Build, test and deploy production AI solutions, turning architecture designs and standards into maintainable, well-tested code.

Design and build agentic AI solutions using orchestration frameworks and Model Context Protocol (MCP), including tool calling, multi-step planning, memory and human-in-the-loop controls.

Integrate large language models, including Anthropic Claude through the Databricks Mosaic AI Gateway and Azure OpenAI, into clinical and operational workflows through secure, documented APIs.

Build on the CDAI platform, consuming governed data products from the Databricks lakehouse through Unity Catalog and exposing AI capabilities through the enterprise API gateway and MCP servers.

Prepare structured and unstructured data, including clinical notes, documents and operational data, for model development, evaluation and retrieval.

Build, train and validate traditional machine learning models, such as classification, forecasting and NLP, where they suit the problem better than Generative AI.

Implement LLMOps and MLOps across the full lifecycle, including versioning of prompts, models and datasets, experiment tracking with MLflow, deployment, observability, cost and latency monitoring, and retraining.

Build prototypes and proofs of concept through the AI Incubator to test the feasibility and value of new ideas, and contribute reusable components to the team's library of AI Accelerators.

Build internal capability by documenting patterns, pairing with colleagues and implementation partners, and supporting divisional staff who use the platform's self-service AI tools.

Things that work for us

• Minimum 5 years' professional experience in software, data or machine learning engineering, including at least 2 years building AI solutions
• Ownership of at least one AI-enabled solution from prototype through production deployment and support.
• Advanced Python engineering skills and working capability in at least one complementary language such as TypeScript/JavaScript, C# or Java, with experience building clean, tested services, APIs and lightweight user-facing applications, plus strong SQL and data wrangling skills.
• Proven experience integrating large language model APIs, such as Anthropic, Azure OpenAI or OpenAI, into applications, including authentication, rate limiting, streaming, structured outputs and error handling.
• Hands-on experience building agentic AI solutions with orchestration frameworks and/or MCP, including tool calling, state management, multi-step workflows, human-in-the-loop controls, failure handling and evaluation.
• Practical experience building RAG solutions with vector databases or indexes, such as Databricks Vector Search, Azure AI Search or pgvector, and embedding models.
• Strong prompt, context and harness (desirable) engineering skills
• Experience across the traditional machine learning lifecycle, including feature engineering, training, validation and deployment, using libraries such as scikit-learn, PyTorch or Hugging Face.
• Experience in technical discovery and workflow analysis, ideally gained in a consulting, implementation or product engineering role working directly with users, would be highly regarded.

Things that work for you:

• Permanent, full-time role based in Melbourne with hybrid working (minimum one day a week on site)
• Competitive salary packaging options
• Discounted private health insurance
• Employee Assistance Program for staff and their families
• Access to the Fitness Passport (FP)

Why become part of St Vincent's Health Australia

Because you're like us, you believe in our vision that every person, whoever and wherever they are, is served with excellent and compassionate care, by a better and fair

Better for all our people, who work tirelessly to provide the highest levels of care and quality of life for our patients and clients. And better support for the local communities we play an important role in and continue to su