MLOps news and analysis on Web Pulse: 3513 curated articles tracking the latest MLOps developments, tools, and research, updated continuously from vetted sources.
FerryAPI has published a template for AI API gateway fallback policies that classifies traffic into five tiers—critical user-facing, non-critical user-facing, internal automation, batch jobs, and experiments—each with it…
A developer spent a month attempting to build a system that could predict multi-agent AI failures before they occur, pre-registering all signals and metrics to avoid confirmation bias. The experiment failed cleanly, with…
FerryAPI has introduced a four-layer cost attribution stack for SaaS teams that call multiple LLM providers, enabling per-request tracking by tenant, feature, assistant, thread, and model. The system uses an OpenAI-compa…
A developer has released LLMhop, a stateless HTTP router that directs OpenAI-compatible API requests to the appropriate LLM inference backend based on the model name in the request body. The single-binary tool, built in …
A developer has outlined a practical method for auditing AI API costs by team and user, addressing the common problem of unexplainable invoice spikes. The approach requires request-level attribution, joining gateway trac…
Citigroup deployed an AI document-processing system that cut the document-review stage for U.S. account openings from roughly one hour to 15 minutes, according to Tim Ryan, the bank's head of technology. Ryan told Reuter…
FerryAPI has published a practical staging checklist for teams evaluating OpenAI-compatible AI API gateways, emphasizing that the gateway should allow apps to switch providers by simply changing the base URL and API key …
Intencion launched a product analytics tool designed specifically for AI agents in production, capturing every user interaction end-to-end including intent, agent steps, and outcomes. The tool installs with one line of c…
A developer built a cost-attribution system for LLM usage by tagging every gateway span with `team.id`, `project.id`, and `feature.id` alongside OpenInference token-count attributes, shipping those spans through an OTel …
Mendral, an AI DevOps agent developer, cannot provide a single accuracy metric for its CI failure diagnosis agent despite processing 36,564 investigations across 5.7 million CI jobs and 14.4 billion log lines in the last…
At Observe 2026, Arize AI launched new Arize AX capabilities designed to automate the agent improvement loop for production systems, including Signal for continuous failure detection, agent orchestration for automated in…
Agentic AI is driving enterprises to prioritize operations platforms as the most critical infrastructure layer, surpassing models and data science talent in importance. The shift to hybrid environments makes operational …
The New Stack published a guide detailing how Azure Kubernetes Service (AKS) can secure AI agent workloads on shared GPU clusters using four control layers: networking, policy enforcement, container image scanning, and r…
Kubernetes security has become more complex with the rise of AI workloads, which introduce new risks around access control, image vulnerabilities, secrets management, and networking. Organizations must adapt their securi…
Depot has launched a general availability release of its CI API and CLI, enabling developers to manage continuous integration workflows entirely from the terminal or editor without needing the web dashboard. The update i…
Sergiu Petean described how his team built a center of excellence to provide SRE as a service, introducing Federated SREs and new roles like production manager and technical tribe lead. They created a culture of data-dri…
A developer has published code and research showing that cross-family evaluation—using a generator and judge from different model families—reduces self-preference bias in AI pipelines by an average of 31.5%. The approach…
AWS Lambda's reported `@maxMemoryUsed` metric tracks a high-water mark across an entire execution environment, not per-invocation memory usage, according to AWS support. A customer's ONNX model inference Lambda function …
Software architects can use "architectural change cases" to anticipate how software systems may need to evolve over time, extending the traditional architecture decision record approach. The method helps teams identify h…
A developer's analysis of LLM fine-tuning versus retrieval-augmented generation (RAG) finds that roughly 70% of production LLM problems are solved by better prompting or RAG, with fine-tuning reserved for the remaining 3…