Securing Agentic Platforms with Zero-Trust Workload Identities

Learn to secure agentic AI platforms using zero-trust workload identities in this comprehensive conference talk from DevConf.IN 2026. Explore the unique security challenges posed by agentic architectures, including dynamic policies, autonomous decision loops, continuous model execution, and cross-service actions. Discover the complete identity flow for securing these complex systems, from attesting compute environments to verifying workload lineage and implementing cryptographic identity solutions with SPIFFE/SPIRE. Master OIDC federation integration and fine-grained authorization enforcement through purpose-built control loops. Examine proven patterns for securing AI agents, vector databases, model-serving pipelines, and GPU/Confidential Compute workloads. Gain insights into critical design patterns, identity lifetime management, trust-domain boundaries, and workload attestation using hardware-backed roots of trust. Build expertise in creating platforms where every component—from operators to model pipelines—authenticates and authorizes seamlessly within a zero-trust security framework.