Research

We secure applications deployed as containers, Kubernetes-managed microservices, and serverless functions. Our research analyzes workload behavior and cross-service execution to enforce least privilege, prevent privilege abuse and policy bypass, and maintain security as applications, dependencies, and permissions evolve.

• Container and Kubernetes runtime security

• Serverless IAM and function-chain security

• Network, file, and cloud-resource access control

• Multi-cloud workload protection and zero-trust enforcement

We design low-overhead mechanisms that observe and constrain behavior at the operating-system, host, and data-plane layers. By leveraging eBPF, LSM, XDP, SmartNICs, and other programmable mechanisms, we secure system calls, file and process activity, and network traffic while preserving system performance.

• eBPF/LSM-based in-kernel monitoring and enforcement

• Stateful and fine-grained system-call filtering

• File-access control and container runtime isolation

• XDP/SmartNIC-based security monitoring and offloading

We transform human intent, program semantics, and runtime evidence into accurate and enforceable security policies. By combining program analysis, NLP, LLMs, and agentic AI with continuous verification, we automate the policy lifecycle—from generation and translation to validation, enforcement, and adaptation—across heterogeneous systems.

• Intent- and context-aware policy generation

• Least-privilege inference and cross-platform policy translation

• Policy conflict, inconsistency, and misconfiguration detection

• Continuous policy verification and lifecycle management

• Access control and security assurance for emerging platforms, including MCP-enabled AI services and XR systems