MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641068269 A) filed by Algoleap Technologies Private Limited on June 01, 2026, for Adaptive Agentic Mesh System And Method For Performing Closed-Loop Adaptation By Iteratively Deploying Updated Configurations.
Inventors include Prashant Achanta; Pani Baruri; Jay Jayakeerthy; Hem Chandra Padhalni; Karthik Nooney; Sairam Chaitanya Ventrapragada; and Srujana Korukoppula.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: ADAPTIVE AGENTIC MESH SYSTEM AND METHOD FOR PERFORMING CLOSED-LOOP ADAPTATION BY ITERATIVELY DEPLOYING UPDATED CONFIGURATIONS Embodiments herein provide a method for an adaptive agentic mesh system (100) to perform a closed-loop adaptation by iteratively deploying updated configurations based on identified performance gaps. The method includes (a) identifying performance gaps, (b) automatically generating and validating updated configurations, (c) automatically deploying the updated configurations, and (d) generating operational data upon deployment of the updated configurations and the operational data is utilized in a subsequent iteration of identifying the performance gaps, thereby creating a continuous improvement cycle with each iteration being performed based on previous optimizations. The adaptive agentic mesh system (100) implements (i) a dual-source reasoning to perform data grounding by prioritizing structured data from a knowledge graph (120) over unstructured data from the retrieval-augmented generation (RAG) (124), and (ii) a principle of least privilege (PoLP) enforcement that provides unique agent identity for each autonomous agent integrated with an identity and access management (IAM). FIGS. 8A-8B
Disclaimer: Curated by HT Syndication.
