The AI Revolution: Where Data Science Goes Next Part 2

Welcome back to Data Science Melbourne! After our last meetup exploring the AI revolution from a range of perspectives, we're continuing the conversation by looking at how AI is changing not just what we build, but how we solve problems, design data systems, and think about the role of data science itself.

This evening, we'll hear from three experienced practitioners who will explore the transformation from complementary perspectives: from how AI has changed the day-to-day practice of data science, to what the rise of AI agents means for the data and ML infrastructure underneath our products, to the bigger questions this shift raises about what data science is — and where it's heading.

Whether you're interested in becoming more effective with AI, building data and ML systems for an AI-first world, or simply trying to separate genuine transformation from the hype, tonight promises plenty of practical insights and some bigger questions to take away.

Talk 1

[Kasturi Deshpande](https://www.linkedin.com/in/kasturi-deshpande-098638156/) is a Data Scientist at Sportsbet, with a background spanning financial services, retail, and analytics consulting. She has worked across pricing, forecasting, and regulatory analytics, with a particular focus on making data science practical and scalable. She also shares her experience through guest lectures at the University of Melbourne and talks at local meetups.

Kasturi will take us through a look back at how AI has changed data science problem-solving — from the cautious early days of adoption to a world where AI is increasingly embedded in scalable, team-wide workflows. Drawing on her experience across financial services, retail, and product, she'll reflect on what has genuinely changed, what hasn't, and how data scientists can adapt and stay relevant as the field continues to evolve.

Talk 2

[Vaishnavi Rajagopal](https://www.linkedin.com/in/vaishnavi-rajagopal/) is a data leader based in Melbourne who works at the intersection of data platforms and ML products. She has led the design and implementation of lakehouse and ML pipelines behind customer data platforms across Databricks, Snowflake, and AWS. She is now focused on building teams that ship reliable data and AI products, while mentoring engineers stepping into data and AI roles.

Traditional ML batch pipelines are built around an important assumption: that we know the questions and use cases in advance. Precomputed churn scores, LTV models, and recommendations work because their grain is fixed and their consumption patterns are known. AI agents challenge that assumption. Users can define cohorts on the fly, ask questions that invert the way an index was originally designed, and explore counterfactuals that no precomputed table can answer. Vaishnavi will explore where traditional data and ML architectures begin to break under these new demands, and what it would take to turn a nightly scoring job into a model that an AI product can query dynamically.

Talk 3

[Boris Savkovic](https://www.linkedin.com/in/boris-s-23969154/) is currently a Data Science and AI Practice Manager at Australia Post. Prior to his current role, he worked across a number of technology companies in Australia, the US and Europe, spanning medical devices, pharma and energy. He has driven data science and AI outcomes in high-innovation, high-performance environments, including two successful Silicon Valley technology companies that went on to significant IPOs. Boris is passionate about using technology to drive meaningful, impactful, viable and sustainable outcomes — and about making that journey with great, collaborative people.

In his talk, inspired by Descartes' famous "I think, therefore I am", Boris will ask what happens when data scientists start rethinking "everything" in the age of ubiquitous AI. What is the same? What is different? What is genuinely new? What is real, and what is hype? Who knows what? What is truth, and where does the real value of…