Red Hat Community

Does your company commercialize / monetize #opensource software? Please take a short LF Research survey to help better understand the #COSS market: sprou.tt/1gyZfTWObUF sprou.tt/1gyZfTWObUF

In open source, transparency is everything! If you use AI for design or code, just own it. Add an "AI-assisted" note to your commit and share your prompts to give reviewers context and keep licensing clear. #OpenSource #AI #Transparency sprou.tt/1WeHXg1bBbk

How does a Python tensor op become a lightning-fast C++ kernel? Christopher Leonard traces the complete PyTorch call stack. Get ready for #PyTorch Conference Europe with a deep dive! red.ht/47sp9vR

A key takeaway on AI and inference from #KubeCon + #CloudNativeCon Europe

At KubeCon Europe, the CNCF and Red Hat contributed the llm‑d framework to the consortium, and the Kubernetes AI Conformance Program tightened its requirements. Stable in‑place pod resizing and workload‑aware scheduling are now mandatory, both critical for running inference engines without restarts and avoiding deadlocks in distributed training. A new Verify Conformance Bot will move beyond self‑assessments, and 31 platforms are already KAR certified. See what the stricter standards mean for production AI: zpr.io/qLGMxdzn9LZc #KubeCon #Kubernetes #AIInference

Eliminate the 'Rego tax'! Here's a dynamic kubernetes policy generator that uses an AI orchestrator & the Model Context Protocol (MCP) to auto-generate context-aware OPA Gatekeeper policies from live cluster data. #Kubernetes #K8s #Security #Automation red.ht/4rM5Y7D

Moving AI agents to production? Ditch the API keys. Learn how we built Zero Trust AI Agents on #Kubernetes with #Kagenti, #SPIFFE, and Istio Ambient mesh. A security-first deep dive. #ZeroTrust #AI #DevOps red.ht/4s3Iews

Good news, #opensource enthusiasts. We are extending the deadline for #cfp submissions until the 8th of March! So you still have one more week to prepare your best ideas and share them with the community at #DevConf_CZ 2026! Submit your proposal: pretalx.devconf.info/devconf-cz-202…

Agents, tools, and APIs need ZT! Learn how to apply #ZeroTrust design principles—like delegated token exchange & continuous verification—to secure your #AgenticAI systems running on #OpenShift. red.ht/4kXpdJD

Triton kernels: Hand-tuned vs. TorchInductor vs. Helion vs. LLM-generated. Our reproducible benchmark reveals the trade-offs in performance, portability, and tuning cost. Check it out! #LLM #GPUkernels #Triton red.ht/4aeQKm7

Secure your high-performance Triton GPU kernels! 📷 New cryptographic signature support for the JIT cache prevents corruption and denial of service attacks in production. red.ht/46q6KiH

Skip the JITters!  Cut AI model cold-start latency by 2x. We're exploring Model Cache Vault (MCV) and OCI image support for Triton/vLLM kernel caches, secured with Sigstore Cosign for trusted, portable performance. #AI #MLOps #DevSecOps #Triton #vLLM  red.ht/4tdo3gW

Moving AI from chatbots to autonomous Cloud-Native Ambient Agents requires a new architecture. This article covers the AgentOps lifecycle on OpenShift, w/event-driven logic, observability, and Zero Trust identity. Read more! #CloudNative #AI #AgenticAI red.ht/4pOq0NC

.@leighgriffin and Ray Carroll just put this article on Spec Driven Development together and is now live on infoq.com/articles/spec-… -- this is a result of our investigations into the technology in our Emerging Technologies team

Check out how to simplify Edge AI Builds with Verified GitHub Actions Patterns! Vance Raiti and Nick Cao show us how their Edge AI Image Pipelines project tackles hardware enablement friction for RHEL bootc images on NVIDIA Jetson. red.ht/3MxP5yy

How do you stop tool sprawl from overwhelming your AI agents? A larger context window isn't the answer—a smarter retrieval process is. Check out Eoghan O'Connor and Kevin Cogan's implementation of Tool2Vec for scalable enterprise AI! red.ht/4iO3i6E

Triton kernel profiling with NVIDIA's Nsight tools? Joseph Groenenboom and Craig Magina show us how! red.ht/49mLFIl

Inferencing at scale: Ron Haberman, Huamin Chen, and other Red Hat OCTO team members share their contributions to the vLLM Semantic Router. red.ht/3WLK87c

Ivan Font and Donald Hunter do a deep dive on a confidential computing approach to AI inference security - have a look! red.ht/3JmLxxT

Guest author Steven Pousty has been hacking on PyTorch, containers, and NVIDIA, and he has some advice for other interested developers - check it out! red.ht/45Dlhrm