DataScience SG

Details
Happy New Year to everyone =). We would like to kickstart this year with a series of talks by AI Scientists from ST Engineering on Physical AI. Starting with an architecture that connects the three pillars of Physical AI (Agentic, Spatial & Robotics), we will explore why successful Agentic AI systems require moving beyond standalone tools to integrate business processes, infrastructure, and ROI-driven strategy for lasting organisational impact. We will also explore the frontiers of Spatial Intelligence and Robotics Intelligence, dissecting the algorithms that allow diverse, heterogeneous robots to coordinate and compete in complex, high-stakes environments.

We are moving our events to Luma! Please subscribe to the DSSG calendar on Luma. You can find it here.

​Agenda
6:30 PM – 7:00 PM Registration & Networking
​7:00 PM – 7:15 PM Introduction to Physical AI @ ST Engineering by Kai Xin
​7:15 PM – 7:35 PM Agentic AI - Why AI Tools Alone Don’t Create Successful AI Systems by Michal
​7:35 PM – 7:55 PM Spatial AI - When World Model Meets Humanoid by Ruofei
​7:55 PM – 8:15 PM Robotics AI - Robust Physical AI by William
​8:15 PM – 8:30 PM Q&A and Closing

Synopsis
Agentic AI - Why AI Tools Alone Don’t Create Successful AI Systems
Today, powerful AI tools are everywhere, yet most organisations still struggle to translate them into real business impact. Why is that? In this talk, Michal shares perspectives from consulting business units, evaluating AI startups, and designing end-to-end AI systems. He explains why many widely adopted tools, despite being technically impressive, fail to meaningfully transform organisations—and what needs to change for AI to deliver lasting value. The discussion moves beyond tools into the broader system: business processes, ROI and outcome measurement, software engineering discipline, self-hosting and infrastructure considerations, and model selection strategy. Attendees will leave with practical tips on what it actually takes to move from AI adoption to AI impact.

Spatial AI - When World Model Meets Humanoid
Dive into the frontier of Spatial Intelligence where generative agent-based simulations meet the reasoning power of LLMs and VLMs. We will discuss the architecture of scalable data factories designed to fuel robust robotics training and bridge the gap between virtual reasoning and physical execution. Join us to explore how these converged technologies are accelerating the development of advanced humanoid applications and embodied AI.

Robotics AI - Robust Physical AI
Uncover why Physical AI’s upcoming challenge is less about making smarter plans and more about executing reliably in the real world. Learn how coordination breaks down at scale and how NEAR Lab focuses on turning research into deployable autonomy.

Speakers
Kai Xin Thia is VP at ST Engineering Group Technology Office (GTO), heading the AI.DA Strategic Technology Centre (STC). AI.DA STC focuses on the research translation of AI, including Physical (Robotics Intelligence, Spatial Intelligence & Systemic Intelligence with Agents), Quantum, Urban Computing & AI Solutions. Kai Xin works at the intersection of data and product innovation, with over a decade of experience driving innovation across finance (London Stock Exchange, DBS), media (Tech in Asia), eCommerce (Lazada-Alibaba), and healthcare (Khoo Teck Puat Hospital). Kai Xin holds an MSc in Computer Science from Georgia Tech.

Michal Polanowski is the Head of Generative AI at AI.DA STC. He builds production-ready AI systems and advises leadership and internal business units on strategic AI adoption. He focuses on transforming cutting-edge research into AI-first workflows that solve real operational problems and deliver tangible, verifiable ROI. With nearly two decades of experience across big data, data mining, data science, deep learning, and now Generative AI, Michal has seen both the hype and the reality of enterprise AI. His approach is deeply pragmatic, grounded in hands-on experience, focused on what works, and uncompromising about measurable outcomes. Michal holds a PhD in Consumer Economics / Market Research from The University of Georgia.

Ruofei Ouyang is the Head of Applied AI at AI.DA STC, specialising in spatial intelligence & simulation AI. He holds a PhD in Computer Science from NUS. His academic expertise lies in decentralised data fusion and distributed computing. He has established a strong research track record, with publications in top-tier venues in artificial intelligence and robotics, including UAI, AAAI, AAMAS, and Autonomous Robots. His professional profile combines deep theoretical knowledge in probabilistic modelling with practical experience in large-scale data science applications.

William Teo is the Robotics AI Research Lead at AI.DA STC NEAR Lab, where he directs research in multi-robot systems and embodied AI applications. He is concurrently pursuing a PhD in Robotics at NUS MARMoT Lab, specialising in multi-agent robot learning. His diverse academic background includes a Master of Science in Computer Science from Georgia Tech and a Master of Engineering in Supply Chain Management from MIT. Currently, he focuses on bridging academia and industry to make robots Smarter Better Faster.