About the author
Sabir Moglad
I build AI agents in production, and I'm the author of Laws of AI Agents.
I'm an engineer building AI agent systems across multiple production codebases. Laws of AI Agents is the field notebook: every law is a heuristic I learned the expensive way and then went looking for the paper or essay that explained why.
The site is deliberately model-agnostic. The laws apply whether you're using Claude, GPT, Gemini, or open-weights, because the failure modes live in the architecture of agent systems, not in any one model.
What I work on
- Multi-agent pipelines for enterprise document processing
- Retrieval-augmented systems with strict citation contracts
- Agent evaluation harnesses and observability tooling
- Prompt engineering, tool design, and scope reduction
Find me
Get the laws in your inbox
Five lessons, one law each, with the source and a real example. Subscribe →