Double Feature: Golden Signals for Data & Data products for AI products

This edition we've got not one but two topics to help us bring the year to rousing finish

• Topic 1: The Golden Signals for Data: An SRE's Guide to Trustworthy Data Platforms
• Topic 2: Data products to supply AI products
• Date: Thursday 27th November 2025 at 6.00 pm - 8.00 pm
• Networking: 6 pm - 6.30 pm and 7.20 - 8 pm
• Presentation Time: 6.30 pm - 7.20 pm
  (includes welcome, presentation 30 mins & community marketplace)
• Location: MYOB, 168 Cremorne Street, Cremorne VIC 3121
• Ticket: Free of cost, however, registrations and RSVP are required!
• Sponsor: MYOB

The Golden Signals for Data: An SRE's Guide to Trustworthy Data Platforms

As engineers, we have a robust, SRE-driven language for microservice reliability. Yet, for data, we often fall back to a simple "Did the pipeline pass?" This talk argues that this is a critical blind spot, as it ignores silent failures like stale, incomplete, or incorrect data.

This session will explore how to adapt SRE's core principles to the unique world of data. We'll introduce a practical framework, inspired by the original "Golden Signals," to create a shared, objective language for data reliability.

Key takeaways:

• Why "pipeline success" is a poor indicator of data health.
• A powerful framework for measuring and communicating data reliability.
• How to define and use data-centric SLOs and error budgets.

About The Speaker
Prashant "Pk" Srivastava is a Distinguished Engineer at Sahaj.ai , specializing in Site Reliability Engineering, MLOps, and DevOps. With a career spanning since 2004 , he is passionate about applying proven SRE principles to solve modern challenges in data reliability and MLOps. His current focus is on designing modern data architectures and building resilient, high-volume data pipelines for high-performance distributed systems.

----------------------------
Data products to supply AI products
In this panel discussion we will be exploring questions as as:

• What are the specific requirements for data products supplying AI products and experiences?
• How do these differ from data products supplying other use cases?
• How do you protect data used by AI systems?
• How do you evolve data governance for AI applications?
• How do you organise people and processes to bring specialised skills to bear at the right points?
• How do you learn and stay current in a fast evolving world?

Meet the Panel:
Elma O'Sullivan-Greene is a Principal Machine Learning Engineer at MYOB, building AI products that help small businesses succeed. Originally from Cork, Ireland, she's spent her career translating complex data into products that delight users and drive meaningful impact—from her early work analysing semiconductor manufacturing data, through healthcare innovation (epileptic seizure prediction, bionic eye vision restoration), to financial services (ANZ, Liberty Financial) and SaaS technology (Culture Amp). She's delivered production ML systems including credit models, classification systems, recommender engines, and MLOps infrastructure for generative AI applications.

Julian Vido is MYOB’s AI Safety Lead. A qualified lawyer, Julian has previously worked as an advisor to Government on proposed AI safety legislation and the design of mandatory AI guardrails. At MYOB, he is working across the business to drive the adoption of AI guardrails in-line with the Australian Government’s Guidance for AI Adoption.

Prajakta Jangle is a DevOps engineer in the AI Products team at MYOB with development experience across the full stack. She brings a strong cloud-infrastructure perspective from her previous work in MYOB’s platform engineering team. She also enjoys contributing to open-source NLP projects.

David Colls is a transformative technology leader, who likes building things and teams to make the world better. He currently leads Data Platforms and Products at MYOB, and is an author of Effective Machine Learning Teams. David previously established the Data & AI practice at Thoughtworks Australia.