GCP & AI Security Architect

• Network/Security

• Remote

• Permanent / Full Time

About the job:
Title: GCP & AI Security Architect
Start Date: Immediate
Position Type: Contract/ Full-time Employment
Location: Remote across USA/ Canada
  
As a GCP & AI Security Architect, you will lead the design, implementation, and governance of security architectures across Google Cloud Platform (GCP). You will lead the design, development, and securing of agentic AI systems. You will be responsible for Agentic Security Architecture: and you will implement lifecycle governance, cryptographically secured agent identity and policy enforcement (e.g., SAGA/TRiSM principles.
  
Required Skills

• Expertise with Security Platforms: Google SCC, Chronicle, Cloud Armor, VPC SC, Forseti, Open Policy Agent (OPA), CSPM tools (Wiz, Prisma)
• Compound AI / Hybrid Inference security : Orchestration across proprietary LLMs, in-house models, APIs into complex data and goal pipelines.
• Agentic Security Architecture: Implement lifecycle governance, cryptographically-secured agent identity and policy enforcement (e.g., SAGA/TRiSM principles).
• Multi-modal Agents security models: Process and fuse visual, audio, and textual inputs using VLMs and perceptual transformers.
• Scalable Orchestration: Blueprint agents with stream/data registries, task planners, built for QoS, latency, cost.
• Ethical & TRiSM Frameworks: Design systems embedding trust, explainability, privacy, compliant with evolving AI regulations
• Expertise with Agentic AI Frameworks & Orchestration Platforms: LangChain, CrewAI, Haystack, AutoGen..etc.
• DevSecOps, Automation and CI/CD security controls
• GCP Compliance & Governance
• AI/ML Security in GCP: Securing Vertex AI and custom models, protecting training/inference data Guarding against prompt injection, model tampering, and data leakage

  Responsibilities

• Architect and implement secure GCP infrastructure and services, including Identity, Network, Compute, Storage, Data, and AI services.
• Architect end-to-end agentic AI pipelines, including modules for perception, goal representation, planning, decision-making, memory, tool usage, and action execution.
• Design multi-agent orchestration layers, managing inter-agent communication, delegation, memory sharing, and task coordination.
• Integrate LLMs (e.g., GPT‑4/4o, Anthropic Claude), retrieval augmented generation, and vector databases to support conversational, research, and reasoning capabilities.
• Build adaptive planning and learning loops, leveraging reinforcement learning, heuristic planning, and continuous feedback to optimize agent behavior.
• Ensure secure, ethical, and trustworthy deployments, including audit logging, provenance tracking, access control, and human-in-the-loop checkpoints.
• Define technical strategies and architecture roadmaps aligned with business goals, overseeing pilot-to-scale implementation and cross-functional stakeholder collaboration.
• Design and maintain GCP security guardrails using Infrastructure-as-Code (IaC) and policy-as-code frameworks (e.g., Terraform, Google Cloud Policy Library
• Lead threat modeling, risk assessment, and secure design reviews for new and existing cloud-native applications.
• Design Zero Trust architectures leveraging BeyondCorp Enterprise, IAM, IAP, and context-aware access.

  Qualification

• A degree in CS, AI/ML, Robotics, Cognitive Systems, or similar.
• Proven track record in R&D or productionizing agentic/multi-agent AI.
• Publications, patents, or contributions to open-source agentic frameworks are a plus.
• Certifications (e.g., ADaSci’s Certified Agentic AI System Architect) are beneficia
• Deep understanding of cloud-native security patterns and regulatory compliance
• Experience with incident response, red teaming, or threat modeling in cloud environments