Harsha Sanjeeva.

I'm a Senior Engineer at Cisco, where I build network access control solutions for Meraki — the platform that manages millions of network devices worldwide. Our services deploy on Kubernetes, which is where my interest in container security started.

On the side, I build tools that answer the question security teams actually need answered: "Given what's running right now, how exposed are we?"

Most Kubernetes security tools stop at scanning. They'll tell you that you have 43,000 CVEs. They won't tell you that 64% of those CVEs are in packages that aren't even loaded in memory. They won't tell you that your ML inference pod has a ServiceAccount with secret access, pod creation, and resource deletion. They won't tell you that your AI agent's delegation chain widens RBAC scope at every hop.

I work on closing that gap. Runtime introspection without eBPF. Blast radius scoring that accounts for network topology. Delegation chain analysis for AI agent workloads. The stuff that matters when you're running a FIPS-hardened, air-gapped cluster where the fancy tools can't deploy.

Here are some technologies I work with:

• Kubernetes
• Go
• Container Security
• RBAC & NetworkPolicy
• AI Agent / MCP
• Runtime Analysis
• eBPF (& alternatives)
• Compliance Frameworks