State of Open Con 2025

AI platform builders have a choice whether to use closed source or open source as the foundation. Closed source platforms provide many aspects of AI platforms, giving you more time to market to build products.

But open source models have really started to catch up, which gives you the ability to build a in-house platform that can be customised and tailored to your needs, with more choice and control.

To build that platform, you need an open source community and a number of components that are yet to be invented. While this emerging landscape already has some powerful capabilities, there's an opportunity for the community to innovate and create new components that can be used to power that PaaS.

As a long time Kubernetes community member, I've been amazed by what the open source community has created. Not only the technology with Kubernetes and the ecosystem of other projects, but also the culture of open source. We have taken what was a potentially limited game to a positive sum game, and now we have the opportunity to do that again in AI.

This panel discussion features engineers and architects from AWS and Mia-Platform building AI applications for gaming and developer experience. They will share their experiences, the latest trends, and where the community can build together to truly make AI for everyone.

Many orgs are evaluating running open source LLMs on their own infrastructure, and Kubernetes is a natural platform choice. However, running open source LLMs in production on Kubernetes is, honestly, a bit of an undocumented mess.

This technical presentation shares the experience of both speakers in deploying production-grade LLM infrastructure on Kubernetes. Through practical demonstrations, we'll explore the complete deployment lifecycle, from GPU setup to optimization techniques like Flash Attention, quantization tradeoffs and GPU sharing.

You'll learn:

* Architectural patterns for efficient LLM deployment using Ollama and vLLM
* Solutions for model weight management and context length optimization
* Techniques for GPU sharing and improving resource utilization
* Production approaches to fine-tuning with Axolotl and serving multiple models with LoRAX

You'll leave with a complete blueprint for building reliable, scalable LLM infrastructure on Kubernetes.

This session explores the importance of building inclusive AI tools tailored for underserved groups, particularly in the Global South. Drawing from JournoTECH's work, we’ll showcase practical steps for bridging the gap despite limited resources and discuss how inclusive design can empower marginalized communities while driving equitable access to technology.

Software is eating the world, and AI is eating software, transforming the way we build and use technology. Open Source has been at the heart of this transformation, ensuring openness and enabling innovation for the past 25 years. However, modern systems increasingly depend on non-code components such as datasets, models, and APIs—creating challenges to completeness that the Open Source must now address.

This session will outline Open Source 2.0: a vision for evolving to global governance and refining—not redefining—the Open Source Definition to ensure that modern systems remain open. We'll discuss how achieving clear community consensus, similar to a constitutional amendment, could ensure that any updates to the definition don't change its meaning for existing systems and avoid the fragmentation of new and conflicting definitions.

By addressing these challenges, Open Source can fully preserve the four essential freedoms of Free Software and thrive in a world shaped by services (cloud) and data (AI).