Cedric Clyburn is a senior developer advocate at Red Hat specializing in open-source AI infrastructure, from efficient model serving to the document pipelines that give enterprise AI applications accurate, usable context. His work applies cloud-native engineering to practical questions of inference costs, data privacy, and reliable retrieval.
Clyburn built his technical foundation around Kubernetes, Linux, containers, and developer tooling. His early work included Buildah, Podman, Tekton, and OpenShift; a developer-advocacy internship at MongoDB expanded his experience with Kubernetes operators and technical education. He subsequently joined Red Hat full-time and moved into a senior advocacy role focused on its AI business. Based in New York, he also helps organize Kubernetes Community Day and has spoken at KubeCon, FOSDEM, DevNexus, and Red Hat Summit.
- Efficient model serving with vLLM. Clyburn teaches a DeepLearning.AI course on vLLM inference, covering quantization, GPU-memory constraints, deployment, throughput, latency, and whether optimization compromises answer quality.
- Private, self-hosted AI infrastructure. His guide to private Model-as-a-Service deployments describes giving developers familiar model APIs while keeping enterprise data, infrastructure, and operating costs under organizational control.
- Layout-aware document processing. His Structuring the Unstructured project uses Docling to preserve tables, columns, captions, images, and section boundaries that simplistic PDF extraction destroys. Local pipelines combine optical character recognition, layout analysis, vision models, and Pydantic representations, producing Markdown, JSON, or HTML without sending sensitive documents elsewhere.
- Chunkless retrieval over document structure. At AI Engineer World’s Fair, Clyburn demonstrated agents navigating a document’s existing outline, selecting relevant sections, and retrieving complete passages without requiring embeddings or vector databases. He also showed how containerized document services and Model Context Protocol integrations make these capabilities available to Kubernetes workloads and AI agents.
His work on enterprise agent architecture extends that operational focus to discoverable tools, policy enforcement, and auditable actions.
