Research Overview
I'm passionate about designing the foundational infrastructure for the Internet of AI Agents — a paradigm where autonomous software agents can interact, negotiate, and collaborate across digital systems without relying on continuous human input. My work explores how to make this transition possible by building decentralized, verifiable, and privacy-preserving protocols for agent discovery, communication, and trust.
Today's internet was designed for static web pages and human-triggered APIs. As we move toward a world with trillions of intelligent agents — coordinating logistics, automating workflows, conducting research — the underlying assumptions of the web must be rethought. We need new abstractions that support dynamic identity, real-time coordination, and programmable trust.
Through Project NANDA, I've coauthored two core papers that examine both the need for — and the architecture of — this new agentic infrastructure.
Beyond DNS: Unlocking the Internet of AI Agents via the NANDA Quilt of Registries and Verified AgentFacts
This paper presents a novel registry architecture tailored for AI agents, addressing the shortcomings of DNS and HTTPS in an agent-dominated internet.
We propose the NANDA Quilt, a federated registry model that:
- Stores only minimal, stable metadata per agent to enable horizontal scalability.
- Links to AgentFacts, cryptographically verifiable documents that describe an agent's endpoints, capabilities, and credentials.
- Enables privacy-preserving queries, dynamic endpoint resolution, and sub-second revocation — all critical at trillion-agent scale.
The architecture separates stable identity from volatile metadata, allowing for:
- Least-disclosure discovery that does not reveal the requester's identity.
- Decentralized trust via verifiable credentials from trusted issuers.
- Multi-endpoint routing, geo-distributed failover, and DDoS resistance.
This system redefines what it means for an AI agent to be discoverable, trustworthy, and interoperable on the open web.
Upgrade or Switch? Do We Need a New Registry Architecture for the Internet of AI Agents?
This companion paper outlines the motivation behind NANDA by evaluating whether current internet infrastructure — including DNS, IP, and certificate authorities — can evolve to support autonomous agents.
It presents a three-stage model:
- 1The reactive web (static HTML/CSS)
- 2Stateless APIs (cloud functions, microservices)
- 3Autonomous agents (stateful, goal-driven, self-coordinating entities)
The shift to agents introduces structural breakpoints: current systems cannot support millisecond-level discovery, instant revocation, or trust frameworks based on behavior rather than domain ownership.
Drawing from historical transitions like dial-up to broadband, the paper argues that the agentic web demands:
- Cryptographically anchored identities decoupled from infrastructure
- Fine-grained, programmable registries
- Execution and discovery models suited for proactive, memory-rich software entities
This is not just a performance upgrade — it is a foundational shift in how the internet handles autonomy, authority, and action.
Ongoing Research Interests
I'm currently exploring the systems and interfaces that will allow the agentic web to emerge as a viable, trustworthy, and human-aligned layer of computation. Specifically:
Agent-Human Collaboration
Developing intuitive, transparent interfaces for interacting with autonomous agents — including multimodal interfaces, holographic projections, and real-time collaboration models that treat agents as collaborators, not tools.
Interface Layers for the Agentic Web
Designing front-end and middleware layers that help individuals and companies expose their services, data, and workflows as agents — effectively "agentifying" websites, APIs, and digital platforms.
Federated Identity and Discovery
Advancing registries that allow agents to be discoverable across organizational, jurisdictional, and protocol boundaries — while preserving privacy and verifiability.
Verifiable Trust and Behavior
Exploring mechanisms for agents to cryptographically prove their capabilities, intentions, and compliance — enabling governance and auditability at scale.
Migration Paths for Legacy Systems
Studying how current web architectures can be extended or restructured to support agentic capabilities without discarding the value of existing platforms.
My overarching goal is to help shape the protocols, abstractions, and tooling required to make the agentic internet not only possible, but inevitable.