System Prompt Design Patterns: Stable, Cacheable, and Composable
Modern system prompts must be cache-friendly and modular. The 2026 system-prompt patterns that ship in production.
8 min read
Agentic AI & LLMs
Agentic AI & LLMs archive page 252 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
Multi-Cloud Claude Deployments: Bedrock, Vertex, and Azure Together
Infrastructure-level look at multi-cloud Claude, including cloud Anthropic strategy, deployment topology, region availability, and cost considerations.
11 min read
Claude Sonnet 4.6 Agent Benchmarks: SWE-bench, TAU-bench, and Beyond
A practical engineering deep dive into Claude Sonnet 4.6 benchmarks, covering architecture, tradeoffs, and what production teams need to know about agent evaluation.
12 min read
Agent CI/CD Patterns: Evals as Tests in 2026 Production Pipelines
Treating evals as the test suite for agents finally clicks in 2026. The CI/CD pattern with PromptFoo, Braintrust, and GitHub Actions that catches regressions before production.
11 min read
AutoGen for Finance Research Analyst Teams: A Production Story
A New York hedge-fund team uses AutoGen 0.5 to coordinate research analyst agents on equities. The team-of-experts pattern in finance and what made it ship.
10 min read
Customer-Support Agent Memory and Conversation History Patterns
Customer-support agents that forget the user's last ticket lose trust fast. The hybrid history + facts memory pattern that retains context across reopened tickets.
10 min read
llama.cpp Server Mode in 2026: CPU-Only Agents That Actually Work
llama.cpp server mode plus quantized models hits real throughput on commodity CPUs. The architecture for CPU-only agent deployments and where this approach makes sense today.
10 min read
Operator 2.0 for Real Estate Search in Miami's Luxury Market
Miami luxury real estate brokers are using ChatGPT Operator 2.0 to search across MLS, off-market networks, and international listings — what is working in 2026.
8 min read
Titans and Long-Term Memory in Neural Networks: Google's Memory-as-Context Work
Google's Titans architecture treats memory as a learnable component that scales beyond context windows. What it does and how it changes long-context design.
8 min read