Data Scientist, Generative AI

About the role

CommBank is hiring a Data Scientist, Gen AI for its Retail Banking Services team in Sydney. You will build, validate and deploy generative and agentic AI solutions across digital and assisted channels, improving customer experiences and helping shape the safe, responsible and scalable use of AI.

Key responsibilities

• Build and evaluate generative AI solutions such as retrieval-augmented generation, conversational assistants and colleague drafting tools
• Design agentic workflows that plan, use approved tools and complete multi-step tasks with permissions and human oversight
• Engineer prompts, system instructions and retrieval pipelines to improve grounding and response quality
• Create evaluation harnesses and golden datasets covering groundedness, accuracy, task completion, latency and cost
• Select and tune models for different use cases, balancing quality, cost, latency and data-residency requirements
• Apply responsible AI controls, including hallucination mitigation, prompt-injection testing, personal-information handling and risk-review evidence
• Work with engineering colleagues to deploy and monitor solutions, including guardrails, telemetry and drift detection

Required profile

• Tertiary qualification in data science, statistics, computer science, engineering or another quantitative discipline, or equivalent experience
• Hands-on data science expertise and comfort combining technical work with clear communication and collaboration across engineering, product, design and governance teams
• Ability to explain model behaviour, limitations and risks to technical, non-technical and governance audiences
• Experience in agile, cross-functional teams with engineers, product owners and designers

Required skills

• Strong Python and SQL skills
• Experience with large language model and agent frameworks, vector stores and cloud AI platforms
• Prompt and context engineering, retrieval design and evaluation of non-deterministic systems
• Delivery of generative AI or machine-learning solutions into production, ideally in a regulated environment
• Knowledge of responsible AI, privacy and model-risk controls, including safety testing and guardrails