Red Hat AI

This is a guide to deploying and benchmarking vLLM with GuideLLM on Kubernetes.

Learn how to fine-tune a RAG model using Feast and Kubeflow Trainer. This guide covers preprocessing and scaling training on Red Hat OpenShift AI.

Learn how to implement retrieval-augmented generation (RAG) with Feast on Red Hat OpenShift AI to create highly efficient and intelligent retrieval systems.

Learn how to implement identity-based tool filtering, OAuth2 Token Exchange, and HashiCorp Vault integration for the MCP Gateway.

Learn how to migrate from Llama Stack’s deprecated Agent APIs to the modern, OpenAI-compatible Responses API without rebuilding from scratch.

Most log lines are noise. Learn how semantic anomaly detection filters out repetitive patterns—even repetitive errors—to surface the genuinely unusual events.

Integrating AutoRound into LLM Compressor delivers higher accuracy for low bit-width quantization, lightweight tuning, and compressed-tensor compatibility.

Optimize AI scheduling. Discover 3 workflows to automate RayCluster lifecycles using KubeRay and Kueue on Red Hat OpenShift AI 3.

Run the latest Mistral Large 3 and Ministral 3 models on vLLM with Red Hat AI, providing day 0 access for immediate experimentation and deployment.

Learn how to optimize AI inference costs with AWS Inferentia and Trainium chips on Red Hat OpenShift using the AWS Neuron Operator.

Use SDG Hub to generate high-quality synthetic data for your AI models. This guide provides a full, copy-pasteable Jupyter Notebook for practitioners.

This performance analysis compares KServe's SLO-driven KEDA autoscaling approach against Knative's concurrency-based autoscaling for vLLM inference.

Learn how to deploy and manage Models-as-a-Service (MaaS) in Red Hat OpenShift AI, including rate limiting for resource protection.

Use the open source SDG Hub to quickly create custom synthetic data pipelines. Train and evaluate your models faster and more efficiently.

Learn how we built a simple, rules-based algorithm to detect oversaturation in LLM performance benchmarks, reducing costs by more than a factor of 2.

Learn how the llm-d project is revolutionizing LLM inference by enabling distributed, efficient, and scalable model serving across Kubernetes clusters.

Learn how we built an algorithm to detect oversaturation in large language model (LLM) benchmarking, saving GPU minutes and reducing costs.

Simplify LLM post-training with the Training Hub library, which provides a common, pythonic interface for running language model post-training algorithms.

Speculators standardizes speculative decoding for large language models, with a unified Hugging Face format, vLLM integration, and more.

Learn why prompt engineering is the most critical and accessible method for customizing large language models.

Oversaturation in LLM benchmarking can lead to wasted machine time and skewed performance metrics. Find out how one Red Hat team tackled the challenge.

Learn how to automatically transfer AI model metadata managed by OpenShift AI into Red Hat Developer Hub’s Software Catalog.

Integrate Red Hat OpenShift Lightspeed with a locally served large language model (LLM) for enhanced assistance within the OpenShift environment.

Explore the benefits of using Kubernetes, Context7, and GitHub MCP servers to diagnose issues, access up-to-date documentation, and interact with repositories.

Dive into LLM post-training methods, from supervised fine-tuning and continual learning to parameter-efficient and reinforcement learning approaches.