Skills: Agentic AI Security specialist

Experience : 7 -12 Years overall technical/security experience (with 1-3+ years relevant hands-on exposure to AI/ML security, LLMs, or agentic frameworks).

Role Descriptions:

Must-Have (Candidates must demonstrate depth in core areas and a combination of the following specialisms)

- Agentic AI Security Engineering: Deep expertise in securing autonomous and goal-based AI agents, multi-agent workflows, and complex tool-calling environments across enterprise ecosystems.
- Agent Identity & Zero-Trust Machine Identity: Practical experience designing and implementing cryptographic, Zero-Trust machine identity for agents using SPIFFE/SPIRE, runtime-bound mTLS certificates, and verifiable runtime attestation to deprecate static credentials and prevent lateral movement.
- Governed Tool & Multi-Hop Authorization: Robust understanding of secure Agent-to-Agent (A2A) and Agent-to-Tool (A2T) protocols, Model Context Protocol (MCP), user-delegated authority (3-Legged OAuth, Token Exchange, nested multi-hop authorization, and fine-grained authorization/downscoping).
- LLM & Agent Vulnerability Assessment: Comprehensive mastery of AI threat taxonomies (OWASP Top 10 for LLMs / Agentic Systems, MITRE ATLAS), specifically addressing prompt injection / indirect prompt injection, tool misuse, unintended action execution, recursion abuse, and data exfiltration.
- AI Red Teaming & Adversarial Simulations: Proven hands-on experience conducting adversarial simulations,



jailbreaking, and red-teaming evaluations on autonomous agents, tool invocations, and RAG pipelines using tools like PyRIT, Garak, PromptFoo, ART, or custom frameworks.
- AI Governance, Guardrails & Human-in-the-Loop Controls: Proven track record in implementing runtime defensive guardrails (e.g., NeMo Guardrails, Llama Guard, Azure AI Content Safety), deterministic parameter verification, action-level human approval gates for privileged operations, and recursion/token budget throttling.
- Security Data Science & Telemetry: Hands-on experience developing telemetry mechanisms, anomaly detection, and end-to-end execution path reconstruction (auditing the full human-to-agent-to-tool invocation chain).
- Programming & Cloud Foundations: Solid proficiency in Python, API security (REST, gRPC, API Gateways), and cloud security principles across Google Cloud (Vertex AI / Agent Engine / Apigee) or Azure / AWS.

Good-to-Have

- Experience with agent orchestration and development frameworks (Google ADK, LangGraph, CrewAI, AutoGen, LlamaIndex, Semantic Kernel).
- Familiarity with AI supply chain security (AIBOM / SBOM, model provenance, artifact signing).
- Knowledge of enterprise AI risk and governance frameworks (NIST AI RMF, ISO/IEC 42001, EU AI Act, PRA/FCA regulatory guidance).
- Professional security certifications (CISSP, CCSP, CISM, Cloud Security Specialty, or specialized AI/ML Security credentials).

📌 Agentic AI Security specialist (Kochi)
🏢 Orcapod Consulting Services
📍 Kochi