Chi Zhang

Chi Zhang trained and worked as an applied artificial-intelligence and machine-learning practitioner before turning his attention to decentralised systems. Public information identifies him as having held senior roles in technology companies including PayPal, where he worked on applied AI problems in product and engineering contexts. His early career experience spans building ML-driven systems that support consumer-facing services and enterprise products, providing a technical foundation for later work that seeks to marry agentic AI with decentralised protocols.

Rather than coming from a purely academic AI background, Zhang’s profile emphasises engineering and product leadership. That practical orientation shaped his approach to platform design: focusing on operational concerns such as scalability, orchestration, and integrating models into real-world workflows. He has spoken publicly about the importance of bridging machine-learning research and production-grade software, a theme that underpins both his technical priorities and his arguments for why decentralised coordination models can be useful for AI tooling.

Zhang founded Kite with the stated aim of creating a decentralised protocol for AI agents that can operate, coordinate and be composed across different environments and services. Kite positions itself as infrastructure to support agentic workflows — autonomous or semi-autonomous pieces of logic that can call external services, route tasks and interact with users — while leveraging decentralised primitives for discovery, incentives and governance. Under Zhang’s leadership, the project has emphasised interoperability, developer tooling and a modular approach intended to let teams assemble agents from reusable components.

In public descriptions of Kite, Zhang frames the protocol as solving practical issues that arise when deploying agentic systems at scale: discoverability of agent capabilities, accountable auditing of agent behaviour, and mechanisms to align incentives between builders, operators and end users. Rather than promising to replace centralised platforms outright, the project promotes a hybrid model where decentralised layers provide coordination and trust properties on top of existing compute and model providers. Zhang’s role combines product strategy, technical direction and fundraising oversight as the project evolves.

Zhang’s stated vision ties together two trends: the rise of increasingly capable AI agents and the demand for decentralised coordination mechanisms that avoid single points of control. He argues that composable agent architectures, paired with transparent economic and governance mechanisms, can enable new classes of applications where agents can be discovered, licensed, combined and audited across organisational boundaries. This vision attends to both technical interoperability — APIs, standards and SDKs — and socio‑technical questions around responsibility and incentives.

In recent activity, Zhang has focused on building the developer ecosystem around Kite, engaging with potential integrators, and refining protocols for agent discovery, reputation and payments. He has also addressed common industry concerns such as safety, reliability and the role of centralised compute providers in a decentralised stack. As with many founders operating at the convergence of AI and web3, Zhang’s work involves both technical development and community building, and the project’s future traction will hinge on developer adoption, integrations with model and compute providers, and how effectively it can operationalise the governance and incentive mechanisms it advocates.