SUMMARY/ABSTRACT/SHORT DESCRIPTION (for publishing)
When people talk about "open source" in today's AI & ML landscape, they likely think of open models and kick off _another_ debate between "open weights" and truly "open source" models. But open source goes beyond models... it's behind the most well-known and critical pieces of the ecosystem, from **training** to **chunking** data & **embedding** for **RAG** to **inference** to **scaling**, **evals**, **observability**, and **optimization**. Open source is _everywhere_ and driving much of the AI ecosystem. For this session, we invite technical practitioners and engineering leaders to join Red Hat and open source expert Wesley Chun to learn what some of these critical pieces are, including but not limited to: PyTorch, vLLM, llm-d, OGX (formerly Llama Stack), Docling, and others.

LONG DESCRIPTION/FULL ABSTRACT (for publishing and/or program committee members)
When you think "open source" in today's AI & ML landscape, people immediately think of open models and kick off yet another debate between "open weights" and truly "open source" models. However for practitioners, a model on its own is useless without the ecosystem to train, feed, serve, and scale it. Driving much of that behind the scenes is open source. This session highlights several foundational open source projects that form the operational reality of AI today. Led by Red Hat open-source expert Wesley Chun, attendees will explore the critical systems doing all the heavylifting in the AI industry today:

* **Training:** How **PyTorch** remains the seminal project empowering users to build their own models from the ground up
* **Data Processing:** Utilizing **Docling** to parse complex documents for vital embedding and vectorization workflows
* **Inference:** Driving high-performance model serving with **vLLM** or **SGLang**
* **Scaling:** Managing complex, multi-server deployments using the **llm-d** Kubernetes platform
* **Applications:** building agents/AI apps with **OGX** (formerly Llama Stack), providing a common agentic API surface in a server; OpenAI-compatible, any model, any infrastructure
* **Managing:** Debugging, evaluating, monitoring, and optimizing with **MLflow**, the biggest open source AI engineering platform for agents, LLMs, and ML models

Whether you are building from scratch or consuming AI services, this talk will illuminate the open source projects you should be aware of and maybe already using every day. As one of the original "OG"s of the open source ecosystem, it may also surprise you to discover which of these communities Red Hat is actively participating in.

LENGTH
This is a 30- to 45-minute technical session talk aimed at AI developers, practioners, and engineering leadership.

AUDIENCE/WHO SHOULD ATTEND and PREREQUISITES [Beginner/Intermediate]
Developers building agents, or other AI-powered web or mobile apps interested in learning about some of the well-known or breakthrough AI open source projects.

TAGS
open source, OSS, OGX, Docling, vLLM, PyTorch, llm-d, LLama Stack, Google, Red Hat, agents, agentic, developers, LLM, LLMs, models, AI, ML, AI/ML, web, desktop, mobile, developer, training, Docling, chunking, embedding, vector databases, vector storage, embeddings, inference, SGLang, Kubernetes, debugging, evals, monitoring, optimizing, MLflow, Meta