Manoj Nair

Manoj Nair is Snyk’s chief technology officer and chief innovation officer, responsible for developing security systems capable of keeping pace with autonomous AI agents. His central argument is that software-writing agents cannot be trusted to certify their own work: secure development requires independent validation, continuous oversight, and defenses embedded directly into engineering workflows.

From enterprise security to agentic defense

Nair earned a master’s degree in computer science from Clemson University and held product and security leadership positions at RSA Security, Dell EMC, and Hewlett Packard Enterprise. At RSA, he oversaw an identity-security portfolio that included SecurID; his subsequent work encompassed public-cloud strategy and converged-cloud software. He holds more than a dozen patents involving information management and security.

He co-founded hybrid-cloud company HyperGrid, serving as chief product officer before becoming chief executive in 2019. He later became Commvault’s chief cloud officer and general manager of Metallic, overseeing its cloud and software-as-a-service businesses.

At Snyk, Nair leads the Emerging Technologies and Solutions Office, which directs incubation and acquisition strategy. He helped explain the company’s 2024 acquisition of application-security specialist Probely, introduced Snyk Labs in 2025, and subsequently outlined Evo by Snyk, a system for coordinating defenses across AI applications and agentic development.

What autonomous software demands from security

• Independent validation for AI-generated software. Nair insists that code generation and security validation remain separate responsibilities. Model-based reasoning can expose subtle weaknesses, but deterministic checks, enforceable policies, and evidence of exploitability provide the independent safeguards enterprises need. His analysis of AI-assisted remediation describes how those capabilities work together.

• Security across the complete agentic attack surface. Agents introduce risks beyond vulnerable code: poisoned skills, compromised Model Context Protocol servers, unsafe dependencies, excessive permissions, prompt injection, and sensitive-data exposure. Nair’s approach examines what an agent can access, how it behaves, and whether its output is safe; Evo Agentic Development Security applies those controls within development workflows.

• AI security posture management. Governance begins with identifying the models, agents, tools, datasets, and libraries already operating across an organization. Nair’s agent-security framework combines continuous discovery with model-specific risk intelligence, behavioral controls, and policy enforcement.

• AI security engineer. Nair sees AI security engineering as a discipline spanning secure architecture, adversarial testing, software supply chains, and autonomous-system oversight. His agentic OODA-loop framework pairs human judgment with defensive agents that observe, assess, decide, and respond as threats evolve.

At the 2026 AI Engineer World’s Fair, Nair connected these ideas to a practical enterprise challenge: establishing whether autonomous agents can be trusted, what risks they introduce, and how to keep their growing output secure.