Physical AI Meetup

Hear 3 expert talks, take the mic for an AI/hardware demo or announcement, and connect over pizza & drinks.

Important: Please register via Eventbrite for your free ticket here – registration is required and spots are limited.

Pizza and drinks for this meetup are made possible by our sponsor, the Curtin Institute for Data Science – Australia's largest university-based data science research institute, applying data science and high-performance computing to real-world problems across government, industry, and academia.

Take the mic at our community open mic.
Share a quick demo, prototype, or announcement – whether it's AI, hardware, or something you're building. All welcome.

Agenda:

5:30 - 6:00 pm | Pizza & Networking
Enjoy pizza and drinks while connecting with other attendees.

6:00 - 6:05 pm | Welcome & Intro
Hear a short welcome from one of the meetup organisers.

6:05 - 6:30 pm | Talk 1: Getting Deep with Neural Networks – Considerations for Physical AI in Autonomous Underwater Vehicles

• Overview: Autonomous Underwater Vehicles (AUVs) are playing an increasingly important role in exploring, monitoring, and defending the world's oceans. However, the underwater environment raises a number of significant challenges for autonomous vehicles; not only must they be physically robust, they must also be smart enough to operate with minimal communication to the outside world. Advanced Navigation's Hydrus, a micro hovering AUV, provides an interesting case study of the hardware selection process for physical AI applications.
• Speaker: Peter Roberts, AI Team Lead, Advanced Navigation
  Peter Roberts is AI Team Lead in Advanced Navigation's Subsea Division. He leads the development of AI-based vision algorithms, which enable the company's micro hovering autonomous underwater vehicle Hydrus to complete a range of autonomous tasks in challenging underwater environments. Peter originally became interested in data driven prediction techniques while working as a chemical engineer on offshore facilities. Through his interests in surfing and photography, Peter has had a long-standing passion for protecting the marine environment. He started work with Advanced Navigation after conducting research into sub-sea structure detection in images using machine learning techniques.

6:30 - 6:55 pm | Talk 2: From Mine Site to Mindset: Real-World Physical AI with Computer Vision (BHP) and LLMs (Monti)

• Overview: This talk explores two applied AI projects delivered end-to-end in real operational contexts. It begins with the BHP Magnetic Plug Computer Vision project, where a vision pipeline was built to detect plugs, assess wear, and turn messy imagery into reliable, decision-ready outputs. It then moves to the Monti LLM project, where an NLP and LLM-driven system was enhanced to support safe, structured conversations aligned with real user needs. Across both case studies, the focus is on what it takes to move from "a model that works" to "a system people can trust and use".
• Speaker: Uno Fang, Data Scientist, Curtin Institute for Data Science
  Dr Uno Fang is a Data Scientist at the Curtin Institute for Data Science, where he works on applied machine learning with a focus on computer vision and data-driven decision support. His work spans end-to-end delivery, from dataset design and model development through to deployment-oriented pipelines and stakeholder-ready outputs. He collaborates with researchers and industry partners on projects involving image and video understanding, sensor-based analytics, and trustworthy evaluation. Alongside industry work, he teaches Business Analytics and enjoys translating technical AI work into practical outcomes.

6:55 - 7:20 pm | Talk 3: Agents for the Bottom Line: How AI Agents Are Starting to Do Real Work

• Overview: AI agents are evolving quickly. With new model capabilities and tool integrations, they can now retrieve knowledge, interact with systems, and support real workflows. But turning an AI prototype into something that reliably operates inside a business is far from trivial. In this talk, Mohinder Jaimangal walks through the design of a real-world AI agent built to operate within a customer-facing business environment. The system retrieves operational knowledge, guides users through processes, and supports workflows that would traditionally require significant manual effort, often consuming hundreds of staff hours. The session explores practical delivery patterns for agent-based systems, lessons learned from deploying AI in production, and where operational AI systems are heading next.
• Speaker: Mohinder Jaimangal, Founder, JupiterAI
  Mohinder Jaimangal is a mechatronics engineer and technology entrepreneur building practical AI systems that automate real operational work. Over the past 20 years he has founded and led multiple technology companies across health, software, and AI. Mo previously built two med-tech companies, including Curve Tomorrow and WeGuide, delivering digital products that improved healthcare outcomes globally. Since exiting in 2024, he is now building his third company, JupiterAI, focused on applied AI and agent-based systems that move AI beyond demos and into real-world operations.

7:20 - 7:30 pm | Community Open Mic
Short informal shares – demo, prototype, or quick announcement.

7:30 pm | Wrap-Up
Any further Q&As and chats can continue here.

Last reminder: A free Eventbrite registration is required to attend – register here.