AI Trust Fabric
The Zero Trust and AI governance architecture I designed for a national-scale insurance data ecosystem — where the thing requesting access to sensitive records may be an autonomous agent rather than a person.
Secure Insurance Data Management Device
Registered Design No. 485611-001, Government of India (Designs Act 2000 / Designs Rules 2001, Class 14-02). Co-registered with Krishan Dev Nidumolu and Yoganand Tadepalli.
An industrial design registration — not a patent — covering the secure insurance data management architecture: Zero Trust principles, adaptive trust evaluation, identity-centric access controls, and secure data access mechanisms.
MCP Guardian
MCP Guardian: A Security-First Layer for Safeguarding MCP-Based AI Systems
Published in Computer Science & Information Technology (CS&IT), Vol.15 No.9, pp.107–121, 2025. DOI: 10.5121/csit.2025.150908
A collaborative research paper (eight co-authors, eight companies) proposing a middleware layer for the Model Context Protocol that adds authentication, per-token rate limiting, WAF-style scanning for malicious tool calls, and request logging — without requiring changes to individual MCP tool servers. The paper’s own evaluation found the layer successfully blocked destructive command patterns and unauthorized-token requests, and measured its overhead at roughly 3–4ms median latency (10–15%) on a reference implementation.