Q&A with Ophir Zahavi at H2O.ai: Running Secure, Automated, Multi-Cloud Infra for AI/ML at Scale

Ophir Zahavi and his team at H2O.ai are building an advanced, automation-heavy, multi-cloud, zero trust AI infrastructure, which supports both customer deployments and internal research environments with a focus on speed, security, and scalability. As the Senior Manager of Cloud Engineering, Ophir leads efforts to ensure that security is seamlessly integrated into the developer experience, using tools like infrastructure as code and automated access controls to enable high developer velocity without compromising on safety. The team employs a zero trust model, emphasizing identity-driven access and automation to manage complex, multi-cloud Kubernetes environments, while also prioritizing a culture of security across globally distributed teams. Ophir stresses the importance of starting with pain points and ensuring that secure pathways are the easiest for developers to use, thereby embedding security into workflows as a default rather than an obstacle. Through these practices, H2O.ai aims to democratize AI by empowering enterprises and public-sector agencies to create specialized AI applications on their private data.