MLOps news and analysis on Web Pulse: 3358 curated articles tracking the latest MLOps developments, tools, and research, updated continuously from vetted sources.
A developer-authored definition of AI observability argues that tracking prompts, model responses, retrieved context, tool calls, and agent decision sequences is necessary because AI systems can return successful HTTP re…
CloudZero launched AI Signals, a cost management platform that gives finance and AI transformation leaders real-time visibility into AI spending. The product is aimed at helping finance teams control the costs of their o…
Researchers introduced SWE-Serve, a benchmark of 53 repository-grounded tasks derived from recent production changes to SGLang, spanning six inference engineering families, to evaluate agents on production inference engi…
A developer built a read-only "AWS Account Investigator" multi-agent crew on Amazon Nova Pro that wires real per-span cost into every trace, reproducing three silent-waste patterns where agent runs complete successfully …
A developer argues that the common Airflow-versus-Dagster orchestration debate misses the real failure mode in ML pipelines: task-centric DAGs treat models and datasets as opaque task outputs rather than versioned, linea…
A developer-authored comparison names Nango as the best platform for building, running, and observing AI agent integrations, arguing that teams can investigate a failed operation and change the integration that produced …
A developer detailed a year-long setup for running coding agents like Claude Code and Codex CLI on real products, built around a single 32-core, 128 GB Linux box reached over Tailscale and managed with a terminal workspa…
A developer building an internal AI assistant chose Amazon Bedrock Managed Knowledge Bases over a self-managed RAG pipeline, delegating document parsing, chunking, embedding generation, indexing, and retrieval to the man…
DeepSeek unveiled DeepSeek Elastic Compute (DSec), a sandbox infrastructure for large-scale AI agent training, in an arXiv paper authored by more than 130 researchers including founder Liang Wenfeng. The paper reports a …
A developer triaged 1,895 open GitHub issues across n8n, Supabase, Cal.com, Appwrite and Home Assistant using a fixed set of nine questions — covering issue kind, reproducibility, version reporting, expected behavior, ac…
Parity engineers ran a self-hosted inference trial of the open-weight DeepSeek V4 Flash model that handled more than 144,000 requests and almost 12.9 billion tokens from 25 engineers between 16 August and 12 September, a…
A developer has detailed a permission-aware retrieval-augmented generation (RAG) pattern that pushes access-control filtering inside the pgvector query rather than applying it after retrieval. The approach, implemented i…
A developer argues that AI agents making real changes to code, configs, and infrastructure leave insufficient evidence behind, since a Git commit alone cannot show what the user asked for, which tools the agent called, w…
A developer argues that the RAG-versus-fine-tuning debate is now a three-way trade-off, with long-context prompting emerging as a legitimate third option as frontier and mid-tier models routinely support 1–2 million toke…
Comet's open-source Opik platform, which provides LLM and agent tracing, evaluation, prompt management and monitoring, has shipped an update, according to the project's public documentation. Opik is Apache-2.0 licensed, …
Langfuse, the open-source LLM engineering platform for tracing, evaluation, prompt management and debugging, was last updated in September 2026 and now carries 35k stars on GitHub. The TypeScript-based project, which ing…
A developer has released Graphiti Local, a tool that runs the Graphiti agent-memory framework on a single embedded file using Ollama and LadybugDB, the maintained fork of the deprecated Kuzu graph database. The setup avo…
Spectro Cloud reported that 85 of its engineers processed 41 billion tokens in a one-month pilot of its PaletteAI Inference Launchpad, with 40 billion tokens handled locally on a single server with eight NVIDIA B200 GPUs…
Uber detailed a feature logging framework it built to keep the features used at inference identical to those used in training, addressing online/offline mismatches such as language values formatted as "en" at serving ver…
A developer's technical breakdown of local LLM memory requirements shows that a 70-billion-parameter model needs roughly 140 GB of VRAM at FP16, 70 GB at INT8, and about 35 GB at 4-bit quantization, meaning even an 80 GB…