#artificial intelligence
May 2026
Learning to Construct Practical Agentic Systems
This paper describes an agent framework that enables designers to enforce modularity in agentic systems, by defining pseudo-tools that call LLMs recursively on a restricted context, and proposes novel learning methods for the agentic components required by this framework, namely pseudo-tools and fixed workflows.
Aditya Kumar, Zhihan Lei, Jerry C. Yan et al.
· arXiv.org · 0 citations