Catching Performance Regressions in AI Agent CI Pipelines
Standard benchmarks miss agent regressions because they grade only final outputs. Trajectory-aware evals in CI catch the 20–40% of regressions that single-turn scoring hides.
11 min read
Agentic AI & LLMs
Agentic AI & LLMs archive page 155 of 631
Agentic AI & LLMs
Agentic AI, LLM engineering, and the models behind modern automation — multi-agent systems, LLM evaluation and comparisons, RAG, fine-tuning, AI infrastructure, security, and production AI engineering.
From the blog
Latest articles
TypeScript Patterns for the Claude Agent SDK — Claude code typescript sdk
Claude code typescript sdk: a practical engineering deep dive into Claude Agent SDK TypeScript, covering architecture, tradeoffs, and what production teams need to know about TS AI development.
9 min read
Claude Sonnet 4.6 Vision Capabilities for Document and Chart Unders...
A practical engineering deep dive into Claude Sonnet 4.6 vision, covering architecture, tradeoffs, and what production teams need to know about multimodal AI.
12 min read
Agno (formerly Phidata): Multimodal Agents the Easy Way in 2026
Agno picks up where Phidata left off and ships first-class multimodal agents. Voice, vision, and tools in fewer than 50 lines of well-typed Python code.
9 min read
Cognee: Knowledge-Graph Memory for Agents — A Getting-Started Guide
Cognee builds and queries a knowledge graph from your unstructured data automatically. A walkthrough from install to your first agent integration in production.
10 min read
WebArena 2.0: Real Browsers, Real Tasks for Browsing Agents Today
WebArena 2.0 brings real-browser tasks and harder evaluation conditions for browsing agents. The benchmark numbers and what they mean for real production browsing builds.
9 min read
Operator 2.0 in Toronto, Paris, and Bangalore: Global Deployment Patterns
How ChatGPT Operator 2.0 deployments differ across Toronto, Paris, and Bangalore — local data laws, language quirks, and regional cost economics in 2026.
10 min read
Where Claude Code for non-technical builders is heading
Where agentic development with Claude Code is heading for non-technical builders — orchestrated agents, reusable skills, and how to prepare now.
7 min read
How to measure success shipping apps with Claude Code
The metrics and leading indicators that prove a non-technical PM is shipping well with Claude Code — plus the vanity metrics to ignore.
7 min read