Blog

Engineering notes on AI agents

Practical articles on building, evaluating, and operating AI agents and AI-powered products.

Diagram of an agent loop with input and output, supported by tracing above and evaluations below.

Featured · AI Agents

Building Production-Ready AI Agents

A practical checklist for moving an AI agent from a convincing demo to a system you can operate: scope, tools, evaluation, observability, and safe failure.

15 September 20262 min readSample article

Read article

AI Engineering12 October 20262 min read

Designing Reliable Multi-Agent Systems

Design principles for systems with several cooperating agents: explicit roles, typed shared state, bounded loops, and traceable hand-offs.

Sample article

Two diagrams: a linear rule-based flow from trigger to condition to action, and an agentic flow with a goal, a plan, and tools in a feedback loop.

Agentic Workflows29 September 20262 min read

AI Agents vs Traditional Automation

When rules and scripts are the right tool, when an agent earns its added complexity, and why most real systems combine both.

Sample article