Gunjan Patel is a director of engineering in the Office of the Chief Product Officer at Palo Alto Networks, developing approaches to enterprise AI that keep software productive, secure, and accountable. His work spans cloud-native infrastructure, large-scale operational governance, and AI-assisted development.
Earlier in his career, Patel worked at Cisco, Ciena, and Tigera on networking, virtualization, container orchestration, and security. A 2017 Santa Clara University alumnus, he later became a cloud architect at Palo Alto Networks, where he championed shift-left security. His cloud-native security checklist maps Kubernetes protections to practical controls, including network policies, restricted permissions, trusted images, and clear operational ownership. He has also participated in GoBridge and organized the Go Remote Meetup.
In 2024, Patel described using BigQuery ML to classify cloud resources across more than 170,000 Google Cloud projects, helping teams identify ownership, environment, and cost-center information. His enterprise AI work has included secure coding-assistant deployments using Sourcegraph Cody and Anthropic models: one AWS deployment reached 2,000 developers within three months and reported average productivity gains of approximately 25 percent.
- Asynchronous AI code review: Patel’s independently developed ghostpilot framework moves intensive code analysis into continuous integration, where models can inspect context, test alternatives, and propose improvements without interrupting developers.
- Developer intent before automation: Clearer variable names and updated comments help models interpret application behavior. Patel’s workflow establishes baseline unit tests, then adds boundary conditions, historical Jira bugs, security policies, and application-specific requirements.
- Context-aware security review: Simulated red-team, developer, and engineering-manager perspectives evaluate potential vulnerabilities against technical constraints and business consequences, helping distinguish serious logic flaws from low-value alerts.
- Human accountability for AI-generated software: AI systems can prioritize findings, suggest fixes, and rerun tests, but humans decide what ships. Patel has extended this concern to least-privilege AI agents, emphasizing access controls and oversight as autonomous coding systems become more capable.
