Data & AI Architecture Melbourne schedule

In this 5-minute networking session, the goal is to connect with three new people. Enjoy the opportunity to expand your network!

• What does a practical blueprint for building AI products look like in real environments?
• How do you move from prototypes to reliable, production-grade AI systems?
• Which architectural decisions matter most for performance, governance, and long-term maintainability?

AI agents can sound capable and still be confidently wrong, especially when they cannot reach the tools and data they need. In this session, InfoTrack's Ajay Kumar explains how the company rebuilt its own data architecture around the Model Context Protocol (MCP), turning agents from chatty assistants into productive operators that can order a title search, pull company data, or run a compliance check, rather than simply describe how to do it. Expect a practical look at what InfoTrack built: a developer sandbox, a governed MCP gateway, and shared design patterns. Ajay will also cover what it took to resource the change internally, and share a live proof point from Tilly, InfoTrack's own MCP-enabled customer facing agent.

• Are we building for hypothetical scale instead of real use cases?
• When does “future-proofing” become wasted effort?
• Simplicity vs flexibility: what actually scales better?
• Are modern architectures solving yesterday’s problems?

 Speakers
Mehul Majethia Former Principal Solution Architect Victoria Police
Ujjwal Batra Tech Lead - Data Platforms and Engineering Carsales
Salah Rubaie Enterprise Architecture and Platform Manager South East Water
Srinivasan Venkatadri Principal Data Consultant Financial Services 

Data quality scores, lineage graphs, and governance frameworks were supposed to solve trust. They didn't, because trust isn't a technical property, it's a human judgment. As agentic systems take on increasingly autonomous decisions, the industry keeps treating "architecturally correct" and "humanly trusted" as the same thing, and the gap between them is widening fast. Drawing on real-world patterns from Informatica and Salesforce, this session argues that human agency isn't a UX afterthought. It's a foundational architectural layer, and this talk shows what it takes to design trust in from the ground up instead of bolting it on after the fact.

• Interpreting privacy obligations and standards through a risk-based lens, rather than as compliance requirements alone
• Understanding why business context, objectives, and data use are critical to identifying and prioritising privacy risk
• Using risk-informed governance to guide practical and proportionate decisions across controls, architecture, and service design

2025 saw rapid efforts to adopt AI with sobering number of failures with AI projects. As we enter 2026, the industry is pivoting from "building AI" to fixing the underlying data problems that cause project failure. This session addresses the shift from hype to practical steps how IT leaders can meet demands for interoperability and governance needed to manage the massive system loads of autonomous AI services. We will translate the core predictions of AI in 2026 into actionable design patterns, moving beyond the race to adopt and toward a strategy of architectural resilience.

This session unveils the Architectural Blueprint for an AI-Ready Enterprise, a framework designed for the era of automation. We’ll dive into the essential pillars of success: real-time event streaming, independent data planes, and context architecture layers. Learn how to build serving layers that move at the speed of AI. 2026 is the year to stop experimenting and start architecting—because your AI is only as powerful as the data infrastructure supporting it.

• Agentic architecture patterns: single-agent and multi-agent designs for natural language data access and workflow execution
• Trust and governance: traceability, evaluation, and control mechanisms for reliable enterprise AI outcomes
• Embedded user experience: bringing agentic workflows into chat interfaces like Teams, Slack, and ChatGPT Enterprise

• Where should you optimise for performance versus cost - and how do you make those trade-offs deliberately?
• What architectural patterns enable flexibility without compromising control, governance, or scalability?
• How do you evaluate and communicate the impact of your architecture decisions in terms the business understands? 

• What architectural shifts are required as data platforms evolve to support real-time and agent-driven AI systems?
• How do you design data and AI systems together to ensure scalability, reliability, and responsible governance from the outset?
• What new pattern are emerging as the boundaries between data engineering and AI engineering blur?

Most agentic AI conversations focus on customer-facing copilots or autonomous decision-making. Behind the scenes, a quieter and more immediately practical use case is emerging: AI coding agents embedded directly in the data engineering lifecycle

This session presents a real, in-flight case study from SACCaN Data & Intelligence, platform integrating cancer care data across South Australia's health system. It walks through the concrete architecture underpinning our rollout.

Health data engineering adds constraints most agentic AI examples never have to face: agents that must never touch raw clinical data, pipelines that cannot introduce new tooling, and mandatory human sign-off on every architecture and privacy-governance decision. We'll cover what we automated, what we deliberately kept human, and the phased rollout we're using to get there. This offers a transferable pattern for introducing agentic AI into regulated, high-stakes data environments.

KEY TAKEAWAYS

• A concrete architecture for embedding AI coding agents into a data engineering pipeline: a working rollout.
• Where to draw the line between what agents should automate and what must stay human, tested against regulated clinical data.

Speakers
Katherine Enderling Program Director SACCaN Data & Intelligence
Priyanka Karimajji Data Engineer SACCaN Data & Intelligence