Beyond simply talking about AI, Cognitivo’s recent enterprise-focused training and hackathon offered Australian software engineers and technical professionals hands-on experience developing AI, locally powered by Gigabyte compact supercomputer – AI Top Atom.
AI adoption is no longer a future milestone on technology roadmaps, it is a daily operational reality for many Australian enterprises. Rather than just adopting turnkey offerings, they are developing and deploying their own solutions to ensure AI sovereignty, cost predictability and unlock its true business potential.
This shift toward local sovereign AI infrastructure is precisely what drove the collaborative effort between Cognitivo, Gigabyte and Nvidia. By equipping local professionals with high-performance, compact Gigabyte AI Top Atom hardware, accelerated by the Nvidia GB10 Grace Blackwell Superchip and the Nvidia AI software stack, the event aimed to demystify complex AI development and accelerate the path from conceptualisation to production-ready deployment within Australian organisations.
Watch the video highlights reel, then read on
Developing local talent
Cognitivo’s highly successful enterprise-focused AI event demonstrates the strong level of interest that Australian organisations have in developing and deploying AI locally.
More than 140 participants attended the AI Demo Day during the opening day, followed by another 60 more enterprise attendees on the following three days, showing the event to be a great success. The recent four-day AI event in Sydney brought together industry, academia and technology partners to give Australian businesses practical experience with local AI development.
Led by Cognitivo, the event was supported by Gigabyte and Nvidia, alongside academic and technology partners including UNSW, Macquarie University and LangChain.
Rather than simply demonstrating the technology, attendees learned directly from AI specialists, worked hands-on with local AI infrastructure, and developed and tested AI workloads using Gigabyte AI Top Atom systems.
Learning open-weight AI
The event included a range of speakers and hands-on AI development sessions, with demonstrations from startups working with “open-weight” AI models like LLaMA, Mistral, Phi, Qwen, DeepSeek and Falcon.
Open-weight AI models are AI models, or "weights," that have their core numerical parameters released publicly so anyone can download and run them. Developers can fine-tune open-weight AI models to adapt them for specific industries.
This approach underpins sovereign AI infrastructure by allowing users to host the model locally on their own hardware or private cloud, ensuring their data does not go off-site. Running a self-hosted model can also avoid per-token API fees.
Building futures together
After the AI demo day, the program shifted to two days of dedicated hands-on AI development training hosted at UNSW Sydney. More than 60 enterprise-level attendees participated, including AI and technology teams from leading Australian enterprises ranging from Banking, Finance, Healthcare and Research.
The training program covered practical AI development topics including working with local models, model evaluation and optimisation, AI agents, clustering and other AI development workflows.
Training staff on AI development is essential for establishing sovereign AI capabilities and underpinning business success, says Alan Hsiao, CEO and Founder of Cognitivo.
“AI is not going to take over our jobs, AI is going to be an amplifier for humanity’s productivity,” Hsiao says.
Echoing Jensen Huang “You may or may not lose your job to AI, but you will definitely lose your job to somebody using AI.”
Official certification
As an officially recognised UNSW course, attendees received a UNSW certificate upon completion of the training. Gigabyte supported the program with more than 20 AI Top Atom systems, allowing attendees to work directly with dedicated local AI infrastructure, rather than being limited to presentations and theory.
The combination of Cognitivo’s AI development expertise, Nvidia’s AI ecosystem and Gigabyte’s AI Top Atom hardware allowed attendees to experience a complete local AI development environment in a practical setting.
Vito Chen, Gigabyte AI Product Manager from Gigabyte headquarters, attended the event and participated in the presentations and technical training.
Hands-on with Gigabyte’s AI Top Atom
Gigabyte had the AI Top Atom on display throughout the event, giving attendees the opportunity to see the system in person and learn more about its capabilities for local AI development.
The AI Top Atom is a compact personal AI supercomputer designed for local AI development and running workloads directly on the desktop. Accelerated by the Nvidia GB10 Grace Blackwell superchip and fifth-generation Tensor Cores, it delivers up to 1 petaFLOP of AI performance. This allows it to support large AI models of up to 200 billion parameters locally.
The compact desktop form factor is designed for local AI development, prototyping, fine-tuning and inference. Gigabyte AI Top Utility software support makes local AI development and deployment more accessible.
The final day of the event was a hands-on hackathon where participants put what they had learned into practice, developing and competing using AI Top Atom compact supercomputers.
The prize for first place was a Gigabyte AI Top Atom, with second place winning a Gigabyte Aero X16 laptop with GeForce RTX 5060 Laptop GPU. Third place took home a Gigabyte GO27Q24G OLED gaming monitor. Gigabyte also gave away a GO27Q24G OLED gaming monitor as a lucky door prize during the AI Demo Day.
