New
RAG vs fine-tuning vs long context: what is working for your use case?
Share the use case, the approach you chose and the results.
Topic
MLOps and inference
75 stories in the last 90 days · 2 community threads
In short
MLOps is the practice of getting machine learning models into production and keeping them healthy: versioning data and models, serving them for inference, monitoring their behaviour and retraining when needed. For buyers, inference cost and latency, and the ability to trace any deployed model back to its data and approvals, tend to decide the platform.
Written by Techarda editors · Updated · How we work
Latest stories
The newest MLOps and inference stories from the last 90 days. Each one opens with our summary and a link to the original source.
Cloud Infrastructure
From training to serving: Reflection’s Beam on CoreWeave
Reflection’s open-weight Beam model on CoreWeave spans training with Weights & Biases Models, SFT in partnership with Axolotl, and Day 0 API access on Serverless Inference in CoreWeave Forge.
Containers & Kubernetes
Day 0 support for Reflection’s Beam on Google Cloud: 5x faster perf with llm-d and NVIDIA Dynamo
Today, Reflection released Beam , its first open-weight model built to give developers, researchers, and enterprises open control and customization for frontier software engineering and long-horizon agentic tasks. Released under a permissive Apache 2.0 license with BF16 , FP8 , and NVFP4 checkpoints available at launch, Beam sets a new standard for inference efficiency in the 500B-parameter class…
AI Agents & Automation
October 2026 LangChain Newsletter: Everything we shipped at Interrupt NYC & London
What we shipped at Interrupt NYC and London: LangSmith AI Search, Engine v2, Trajectories, Custom Apps, Fine-Tuning, and Managed Deep Agents v0.9.
Cloud Infrastructure
DigitalOcean Inference Router Enhancements: Performance Tuned Routing, No Rules to Write
Every few weeks, a new model steps up and claims best-in-class at something. The launch chart makes the case, the link makes the rounds, and your team is back to the question nobody has time for: should we switch? The answer takes a week of evals, side-by-side tests and debate, and by the time your team lands on one, the next contender is already on the scoreboard. It’s a race you can’t win…
Cloud Infrastructure
Amazon EC2 G7 instances now available in Europe (Frankfurt) and Europe (Stockholm) Regions
Amazon Elastic Compute Cloud (Amazon EC2) G7 instances powered by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs are now available in Europe (Frankfurt) and Europe (Stockholm) Regions. G7 instances deliver up to 4.6x AI inference performance and up to 2.1 graphics performance compared to G6 instances. G7 instances also deliver faster performance for GPU-accelerated data analytics workloads.…
Cloud Infrastructure
AWS Lambda announces Lambda Web Functions
AWS Lambda announces Lambda Web Functions, a new serverless compute primitive where your code handles HTTP requests directly. Your code listens on a port and handles requests as it would on any HTTP server. You use frameworks you already know, such as Express or Hono, with no adapters or custom configuration. This enables you to serve workloads such as AI inference endpoints, server-side rendered…
Cloud Infrastructure
Deep dive into Azure Maia 300 architecture
Maia 300 operates as a single platform to meet the demands of cloud-scale agentic inference.
Security
Introducing Clef-omni with full multimodality, plus a faster Clef and a cheaper Clef-flash
We are expanding the Clef decision model family with Clef-omni, natively processing audio, video, images, and text in a single pipeline. We’ve also lowered Clef-flash pricing and boosted Clef inference speeds by up to 2.0x.
Earlier in the last 90 days
- OpenAI GPT-6.1 Sol now supports Ultrafast mode on Amazon BedrockAmazon Web Services · 9 Oct 2026
- AWS Cost Explorer, Budgets, and Dashboards now support Amazon Bedrock product attributesAmazon Web Services · 9 Oct 2026
- The Architecture Behind More Efficient AI InferenceEverpure · 9 Oct 2026
- MLOps at the edge: Running AI models at the edgeRed Hat · 8 Oct 2026
- MLOps at the edge: Running AI models at the edgeRed Hat · 8 Oct 2026
- Building an evidence-grounded agentic security operations harness on CloudflareCloudflare · 8 Oct 2026
- HPE accelerates enterprise AI innovation with next-generation AMD CPU-powered serversHPE · 8 Oct 2026
- Anyscale on Azure is Generally Available: Enabling Enterprises to Own the Full AI Loop, Not Just InferenceAnyscale · 7 Oct 2026
- Networking for AI inference model serving - GKE only and for all other backendsGoogle Cloud · 6 Oct 2026
- Closing the Loop Between Inference and Post-TrainingCoreWeave · 6 Oct 2026
- AI cost management: The complete guide for enterprise ITFlexera · 6 Oct 2026
- MLOps for edge AI: Preparing AI models for deployment at the edgeRed Hat · 5 Oct 2026
- Introducing Web Search API via AI GatewayCloudflare · 2 Oct 2026
- Cartridges: A Cache You Train, Not a Context You FillEverpure · 2 Oct 2026
- Key metrics for monitoring DatabricksDatadog · 2 Oct 2026
- How NVIDIA GPUs Help Accelerate OpenAI’s GPT-6 Astra UltrafastNVIDIA · 2 Oct 2026
- Introducing Clef: our open-source decision models, and new RL fine-tuning platformCloudflare · 2 Oct 2026
- Introducing Agent Droplets: everything an agent needs, one price, one billDigitalOcean · 1 Oct 2026
- Equinix at NVIDIA GTC Berlin 2026Equinix · 1 Oct 2026
- OpenAI GPT-6 Astra now supports UltraFast mode on Amazon BedrockAmazon Web Services · 1 Oct 2026
- Amazon Bedrock expands Claude models in-region support in the UK (London)Amazon Web Services · 1 Oct 2026
- Red Hat delivers peak performance on Kubernetes and CPUs in MLPerf Inference v6.1Red Hat · 29 Sept 2026
- Announcing Claude Sonnet 5.5 on Snowflake Cortex AISnowflake · 29 Sept 2026
- Grok 4.7 is now available on Amazon BedrockAmazon Web Services · 29 Sept 2026
- GPT-6 Sol and Luna on Snowflake Cortex AISnowflake · 29 Sept 2026
- How to Integrate LMCache with FlashBlade for Faster LLM InferenceEverpure · 28 Sept 2026
- The agent-first cloud: why we built Managed AgentsDigitalOcean · 26 Sept 2026
- Best practices guide for customizing Gemini models via Reinforcement Learning (RL)Google Cloud · 26 Sept 2026
- New in LangSmith: Engine v2, Managed Deep Agents, Fine-Tuning, and moreLangChain · 25 Sept 2026
- Idle GPUs and runaway token bills: the real economics of production AI - Spectro CloudSpectro Cloud · 25 Sept 2026
- Scribd, Inc. classifies more than 400 million documents with Gemini batch inference on Gemini EnterpriseGoogle Cloud · 25 Sept 2026
- Introducing LangSmith Fine-TuningLangChain · 25 Sept 2026
- Amazon SageMaker HyperPod Inference Gateway for scalable LLM inferenceAmazon Web Services · 24 Sept 2026
- Fine tune your own custom LLM with Canonical Charmed Kubeflow and FeastCanonical · 24 Sept 2026
- 41 billion tokens later: dogfooding local inference routingSpectro Cloud · 23 Sept 2026
- Introducing DigitalOcean Managed Agents: One AI-native stack to power your intelligenceDigitalOcean · 23 Sept 2026
- Connect Amazon Bedrock to PaletteAI Inference LaunchpadSpectro Cloud · 22 Sept 2026
- MLOps Solutions for Production Machine LearningLaunchDarkly · 22 Sept 2026
- What is hybrid AI? The seven-stop spectrum (part 1) | Spectro CloudSpectro Cloud · 21 Sept 2026
- CoreWeave Leads Cloud Providers in MLPerf® Inference v6.1 Performance with NVIDIA Blackwell UltraCoreWeave · 17 Sept 2026
- MLPerf Inference v6.1: pioneering agent, VLM benchmarksLambda · 17 Sept 2026
- NVIDIA Vera Rubin NVL72 Delivers Leading Performance in MLPerf Inference v6.1 DebutNVIDIA · 17 Sept 2026
- AI Infra summit 2026: Enterprise AI takeaways | Spectro CloudSpectro Cloud · 16 Sept 2026
- MLOps Solutions for Production Machine LearningLaunchDarkly · 16 Sept 2026
- Accelerate OpenFold3 with AMD and VultrVultr · 15 Sept 2026
- MLOps pipeline: Stages, tools, and deployment workflowLaunchDarkly · 14 Sept 2026
- Why Agentic Inference Needs Prefix-Aware Routing InfrastructureCoreWeave · 13 Sept 2026
- Agentic Inference in Production: The Four Infrastructure Decisions That MatterCoreWeave · 13 Sept 2026
- Why Open Models Can Win on Enterprise Cost and ControlCoreWeave · 13 Sept 2026
- 5 Lessons from Building a Multi-plane Network Fabric for Agentic AICoreWeave · 13 Sept 2026
- Get Started with Kimi K3 on CoreWeave Dedicated InferenceCoreWeave · 13 Sept 2026
- Where AI Inference Needs to RunEquinix · 10 Sept 2026
Showing the newest 60 of 75. The feed archive lists every story by month.
What practitioners are discussing
Stories say what companies announced. These threads are where people who work with MLOps and inference compare notes.
New
Where do most of your inference workloads run today?
Pulse check for FY27 planning conversations.
Companies in these stories
The active companies with the most MLOps and inference stories in the last 90 days. It counts news coverage, not quality, market share or fit for your needs.
Related categories
Where MLOps and inference sits among the categories Techarda covers. Each hub shows who is active, and each guide sets out how to evaluate vendors.
Terms to know
Plain definitions of terms that come up in MLOps and inference stories, from the Techarda glossary.
Frequently asked questions
What is the difference between training and inference?
Training builds or adjusts a model from data and happens occasionally. Inference is running the model to get answers, which happens every time someone uses it. For most production systems, inference is where the ongoing cost sits.
Does MLOps still matter for generative AI?
Yes. Prompts, retrieval settings, model versions and evaluation sets all change and need the same versioning, testing and monitoring as classic models. Some teams call this LLMOps, but the discipline is the same.
What should we monitor once a model is live?
Latency, errors and cost per request, plus quality signals: drift in the input data, shifts in outputs and, where you can, results checked against what actually happened.
Stories are filed under topics automatically from their headline and summary, so the odd one may sit in the wrong place. Spotted one, or a topic we should add? Email engage@techarda.com. All topics.