Eduardo Fraga (EF-TECH)

.

[Eduardo Fraga (EF-TECH) @eftechcombr]

Gateway API v1.6 — TCPRoute e UDPRoute tornam-se Standard https://t.co/Ih7RoVnXGD

[Kubernetes @kubernetesio]

Gateway API v1.6: TCPRoute and UDPRoute Graduate to Standard - https://t.co/fvQ0MkdQ6i #Kubernetes

[Eduardo Fraga (EF-TECH) @eftechcombr]

Ansible CLI GitHub Action — Run Ansible Commands in Your CI/CD Pipeline https://t.co/6o80dwKnAw

[Eduardo Fraga (EF-TECH) @eftechcombr]

OpenAI Confirma que Modelos de IA Fugiram do Sandbox — Skynet Está Começando! https://t.co/1ykClxWelf

[Kimi.ai @Kimi_Moonshot]

Introducing Kimi K3: Open Frontier Intelligence

🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows

Kimi K3 is now live on on https://t.co/zrk6zZxZUo, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.

🔗 API: https://t.co/XCrgjXAqMw
🔗 Tech blog: https://t.co/YTfiMSNM1f

[Eduardo Fraga (EF-TECH) @eftechcombr]

Argo CD 3.5 Ships a Dedicated ApplicationSet UI: Managing GitOps at Scale https://t.co/hw8RCY6HXC

[Eduardo Fraga (EF-TECH) @eftechcombr]

MCP-GLPI: integre assistentes de IA com o GLPI via Model Context Protocol https://t.co/HBlbqqWz2b

[Eduardo Fraga (EF-TECH) @eftechcombr]

Entendendo a RFC 10008: O Método HTTP QUERY https://t.co/4XFMbFi1fI

[Eduardo Fraga (EF-TECH) @eftechcombr]

Princípios SRE do Google: Evolução da Operações Tradicionais para a Engenharia de Confiabilidade de Sites https://t.co/RXBNTTURwH

[Eduardo Fraga (EF-TECH) @eftechcombr]

Grafana 13.1: Novidades em observability as code, Grafana Assistant e mais https://t.co/iTqKfWmJJj

[Eduardo Fraga (EF-TECH) @eftechcombr]

GLPI 11.0.8 e 10.0.26: imagens Docker e Helm Chart atualizados https://t.co/xloVvYkbha

[Eduardo Fraga (EF-TECH) @eftechcombr]

AWS Summit São Paulo 2026: o maior evento de cloud da América Latina https://t.co/27z3dXr3wR

[Eduardo Fraga (EF-TECH) @eftechcombr]

Como instalar o Hermes Agent usando Kubernetes https://t.co/lm3Bfr0EqO

[DevopsCube @devopscube]

DevOps to MLOps Series 🚀

We have published 8 editions of the MLOps for DevOps series so far.

It is a hands-on, project-based guide to MLOps built specifically for DevOps engineers.

Here is the key idea 👇

Most MLOps resources are written for data scientists learning infrastructure. This series flips that.

You start by building the ML foundation on your local machine - from raw data to a trained, tested model.

Then you move to the infrastructure side.

- Orchestration
- Kubeflow Pipelines
- Feature stores
- Model registries and more..

Every edition comes with working code in a GitHub repo. You can follow along and build as you learn.

𝗚𝗶𝘁𝗛𝘂𝗯 𝗿𝗲𝗽𝗼: https://t.co/bWermJeKNM

If you have feedback, raise an issue in the repo.

We are happy to discuss and improve it.

A quick note on the goal.

This series is built to help DevOps engineers understand the infrastructure side of MLOps with just enough ML knowledge.

This is not an AI/ML course.

♻️ Repost if you know a DevOps engineer looking to expand into MLOps.

#mlopsfordevops #mlops #devops

[Kubernetes @kubernetesio]

Kubernetes v1.36: In-Place Vertical Scaling for Pod-Level Resources Graduates to Beta - https://t.co/bW8hutGKRA #Kubernetes

[vLLM @vllm_project]

🎉 Day-0 support for @deepseek_ai V4 Pro and Flash on vLLM — a new generation of DeepSeek model, purpose-built for tasks up to 1M tokens. Alongside the release, we're publishing a first-principles walkthrough of the new long-context attention and how we implemented it in vLLM.

The new attention mechanism, in four moves:
• Shared K/V + inverse RoPE → 2× memory savings
• c4a / c128a KV compression → 4×–128× savings
• DeepSeek Sparse Attention over compressed tokens
• Short sliding window for locality across compression boundaries

At 1M context, per-layer KV state is ~8.7× smaller than a DeepSeek V3.2-style 61-layer stack (9.62 GiB vs 83.9 GiB, bf16). fp8 attention cache + fp4 indexer cache shrink it further.

vLLM side:
• Unified hybrid KV cache — single logical block size (256 native positions) across all compression rates; compressor state folded into the SWA KV cache spec so prefix caching, disagg prefill, CUDA graphs and MTP reuse the same abstraction
• Three page-size buckets for the full 5-way cache stack → no cross-kind fragmentation
• Fused kernels: compressor + RMSNorm + RoPE + cache insert (1.4–3×), inverse RoPE + fp8 quant (2–3×), Q-norm + KV RoPE + K insert (10–20×)
• Multi-stream overlap of indexer vs main-KV compression vs SWA insertion

Disaggregated serving is supported out of the box and strongly recommended for best performance.

Follow our recipes site for verified commands for @nvidia Blackwell (B200, B300, GB200, GB300) and Hopper (H100/H200/H20) systems.

Thanks to the @deepseek_ai team for open-sourcing DeepSeek V4, and to @inferact for landing day-0 support 🤝

📝 Blog: https://t.co/Eh7vk6xVJy
📖 Recipes: https://t.co/jlWuzYyZeX
🤗 https://t.co/IA9qAysqJk

[Grafana @grafana]

Grafana 13 ships Git Sync as GA!

Save dashboards directly to GitHub, GitLab, or Bitbucket, open a PR from the Save button, and sync changes back automatically.

https://t.co/nh2e3xzeu9

[Kubernetes @kubernetesio]

Gateway API v1.5: Moving features to Stable - https://t.co/qePktu1KCb #Kubernetes

[Eduardo Fraga (EF-TECH) @eftechcombr]

Dominando o OpenCode https://t.co/W9Urj7nH8A via @YouTube