
< keynote />
Fri, April 30Architecture & Distributed SystemsAI-Native SoftwarePlatform Engineering & DevOps
The current AI narrative often suggests that bigger is better: trillion-parameter models running in remote data centres, consuming enormous amounts of power, and charging per token. But architectural elegance is not necessarily found in scale alone. It can also emerge from the balance of passion, purpose, vocation, and utility, a concept the Japanese call ikigai.
In this keynote, you will explore an alternative to relying on increasingly large, remote AI models: personal AI sovereignty through local compute. You will examine how a deliberate, right-sized topology of local machines can help each server and AI model fulfil its potential, rather than simply pursuing the largest configuration your budget allows.
The session explores a pragmatic approach to local AI architecture, including how to match model quantisation to different hardware tiers and design systems around their intended purpose. You will consider how developers can reclaim local compute and build AI environments that are sustainable, appropriately sized, and under their own control.
What You Will Learn
Why bigger AI models and more powerful hardware are not always the right architectural choice
How to design a right-sized topology of local machines for AI workloads
How to match model quantisation to appropriate hardware tiers
How local compute can support a more sustainable, purpose-built approach to AI architecture
Who Should Attend
Software Developers, AI Engineers, Software Architects, Infrastructure Engineers, and technical practitioners interested in local AI, model deployment, hardware optimisation, and personal AI sovereignty.
< speaker_info />
Scott Davis is a Web Architect and Digital Accessibility Advocate, focusing on the multisensory aspects of web development. In a world where half of all Google searches are done by voice, and 80% of all social media videos are watched with the sound off and closed captions on, accessibility is a springboard for innovation.