Open Source Summit Korea 2026

The era of open weights has arrived — and it is reshaping every layer of the AI stack, from chips and infrastructure all the way up to the applications that create real-world value.
This transformation has deep roots. Over two decades ago, NVIDIA opened up GPU computing with CUDA, laying the foundation for the developer ecosystem that would eventually power the AI revolution. That same philosophy continues today — through the Nemotron model family, NeMo Agent Toolkit, and Cosmos — contributing to a growing open stack that any developer, researcher, or startup can build on.
We are now entering the era of agentic AI, where open-source stacks enable developers to go beyond inference and build systems that reason, plan, and act. The tools are open. The models are open. The opportunity is now.
This session is an invitation — to understand the open-source AI landscape, to see where it's headed, and to be part of what comes next.

Zephyr initially set out to solve a problem that many embedded teams quietly struggled with: how to build dependable real-time systems without being locked into a single vendor, toolchain, or proprietary stack. The project introduced a new model built around portability, adoption of open source and security best practices, modern tooling, and a shared ecosystem of drivers and middleware.

From the start, there was the commitment from the start apply known best practices to its development. While Zephyr is a different code base, a lot of the lessons learned from developing the Linux Kernel were applied. The project has also focused on incorporating security best practices from the start which now enables it to make compliance easier for manufacturers looking to conform to the emerging Cybersecurity Resilence Act (CRA).

Best practices have also enabled the project to work towards achieving formal safety certification for 61508 and 26262. The project has achieved 61508 concept approval at this point, as is working towards formal certification, using a combination of traditional V-Model analysis, and innovative techniques to keep up with the speed of open source development.

As AI agents become active participants in software development, organizations need to rethink how they approach open source strategy. This talk explores why Agentic AI introduces new responsibilities for Open Source Program Office (OSPO) teams: from governance and policy to trust, security, and organizational enablement. This session shares a strategic perspective on how OSPOs can help organizations adopt AI agents responsibly while strengthening open source practices

This presentation examines the evolving role of open data in the next phase of artificial intelligence development, and the growing imperative to align innovation with trust. As AI systems become increasingly data-driven, the quality, origin, and governance of data are emerging as foundational issues that will shape not only technological progress, but also public confidence and global competitiveness.
The talk situates these developments within a broader shift toward greater transparency and accountability in AI, reflected in emerging policy directions such as the EU AI Act and Korea’s Korean AI Basic Act. Rather than viewing openness as a risk to be constrained, it explores how responsible data practices and shared governance approaches can enable wider participation in AI development while maintaining integrity and trust.
Ultimately, the presentation argues that the future of AI will depend not only on model performance, but on the strength of the underlying data ecosystem. Building sustainable trust around open data—through transparency, collaboration, and accountability—will be critical to unlocking the full potential of AI at scale.

AI systems need more than intelligence; they need context that persists. Without it, even strong models can misinterpret information, lose decision rationale, or repeat the same mistakes. Context Graphs have emerged as a practical pattern for agentic AI: a living graph that captures not only what was retrieved or known, but how context led to actions through tool calls, constraints, policies, and outcomes, stitched across entities and time so precedent becomes searchable.

This talk explores context engineering as the discipline of designing that context layer, and shows how context graphs complement retrieval by enabling multi-hop, structured context assembly (building on GraphRAG-style hierarchical summaries) while improving explainability and evaluation. Attendees will leave with a practical understanding of how to build context pipelines that combine contextual retrieval with persistent memory and provenance, and why context graphs are becoming central to trustworthy, enterprise-ready AI systems.