MLOps news and analysis on Web Pulse: 3309 curated articles tracking the latest MLOps developments, tools, and research, updated continuously from vetted sources.
Anyscale on Azure became generally available on October 7, 2026, letting enterprises run the full AI loop — data processing, training, post-training, evaluation, and inference — on Azure Kubernetes Service in their own s…
Average GPU utilization across Kubernetes clusters measured in Cast AI's 2026 State of Kubernetes Optimization Report was just 5%, and only 2% on Azure Kubernetes Service (AKS), meaning 98 cents of every dollar spent on …
AGO AI Quality Gate (AGO) is an evidence-first quality-gate framework for retrieval-augmented generation (RAG) systems that enterprises can use to decide whether to promote, revise, or block a system version when evidenc…
The blog post "Streaming an LLM chat across a TypeScript/Python boundary: the details that bite" examines the integration seams that break when streaming LLM chat responses between a TypeScript frontend and a Python back…
MoteDB, an embedded database for embodied AI, fixed three bugs that made a top-k time-series query on INT timestamp columns 680x slower, cutting latency from 1,196ms to 1.76ms on a 1M-row table. The bugs included a top-k…
Frank Guan, writing for Google AI, published a guide on building production evaluation pipelines for AI agents in about 60 minutes, arguing that developers should stop relying on informal "vibe checking" to judge agent b…
A developer built TouchGrass AI, an open-source micro-expedition app that turns short breaks into outdoor quests and uses a local MobileNetV4 ONNX model running on CPU in about 12 milliseconds to classify a single field …
Meta has open-sourced Rebalancer, a C++ and Python library it has used for over 9 years to place shards, servers and traffic, handling about 40 million assignment problems a day. The library uses local search or MIP solv…
A developer argues that AI agents should be treated as proposing actions rather than executing them directly, with a separate control layer handling validation, authorization, and execution. The proposed architecture rou…
OWASP's Q1 2026 exploit round-up documents eight real-world AI agent security incidents, including a three-hour PyPI backdoor of the LiteLLM gateway that pulled an autonomous attack bot into roughly 47,000 downloads, a p…
A developer described building a thin abstraction layer between their application and multiple LLM providers after finding API keys scattered across three places, duplicated retry logic, and no visibility into per-featur…
A developer built a Claude Code plugin called maddog that routes each step of a task to a subagent on a model chosen by the kind of decision the step requires rather than by task difficulty. In 49 advisor-mode sessions c…
A developer has released ccdeck, an open-source local dashboard that visualizes every Claude Code and Codex session on a machine as a live tree, surfacing sessions waiting on input, subagent and tool-call activity, and c…
An arXiv industry experience report (arXiv:2610.06890v1) documents three years of running an event-driven cloud infrastructure for continuous machine learning training in automotive manufacturing, reporting a 72-78% cost…
FluidPD, a prefill-decode disaggregated LLM serving system, improves overall SLO attainment over static SGLang by up to 94.6 percentage points across production Azure trace workloads, according to the arXiv paper 2610.06…
AegisFlow, a multi-agent framework described in arXiv paper 2610.06971v1, cut Mean Time to Repair for fragile data pipelines by 98.1 percent, from an average of 170 minutes per patch to 3.2 minutes, with a 92 percent pat…
A developer building the Second-Memory semantic search system decoupled memory creation from embedding generation using a transactional outbox pattern, a queue, and a dedicated embedding worker writing to pgvector. Memor…
A developer outlined a production pattern for LLM API cost monitoring in .NET, using a DelegatingHandler to capture usage.total_tokens and push metrics to Prometheus with a background reconciliation job, keeping added la…
A developer has published the sixth and final part of a .NET agentic-systems series, detailing a safety layer that redacts secrets before they reach storage rather than at presentation time. The design routes every Conve…
A developer building Second-Memory, a service that stores user-written memories over time, chose PostgreSQL with the pgvector extension over Pinecone to add semantic search to the system's V1 architecture. The engineer c…