ICT Architect

Project overview

For deployment with our client in Zurich, we are looking for an ICT Architect

Responsibilities

• Security Architecture for AI Use Cases:
• Conception of a holistic security architecture for chat-based interactions, RAG applications, agentic workflows, and the use of Foundation Models – including consideration of new attack surfaces and trust boundaries between users, models, agents, tools, data sources, and external services.
• AI-specific Threat Models & Controls:
• Development of threat models and protective mechanisms against, among others, Prompt Injection (direct/indirect), Jailbreaking, Model Manipulation, Data Poisoning, Insecure Output Handling, security-relevant hallucinations, and misuse of autonomous agents; derivation of preventive and detective controls.
• Security-by-Design for RAG & Data Access:
• Security concepts for retrieval processes, vector databases, and connected knowledge sources with a focus on preventing unauthorized information aggregation, data leakage, cross-user/cross-tenant leakage, and enforcement of existing access rights in the AI context.
• Identity, Authorization & Trust Concepts for AI Agents:
• Design of identities, roles, and permissions for users, models, agents, and tools (Least Privilege, delegated permissions, secure agent-to-tool/agent-to-agent communication, control of autonomous actions and critical transactions).
• Guardrails & AI-specific Security Controls:
• Conception of reusable platform mechanisms for input/output validation, prompt/context protection, content filtering, policy enforcement, tool/plugin approvals, secrets handling, sandboxing, as well as restriction and monitoring of agent capabilities.

Requirements

• Must: Several years of experience as a Security Architect (ideally with a focus on cloud/application security) and a solid technical understanding of modern AI platforms, LLM/RAG architectures, and agentic workflows.
• Must: Demonstrable experience in threat modeling and the implementation of security-by-design concepts, including the development of guardrails, access and authorization concepts, as well as monitoring/logging approaches for complex applications.
• Must: Ability to recognize AI-specific risks (e.g., Prompt Injection, hallucinations with security implications, data leakage), translate them into architectural and control requirements, and represent them to both technical and non-technical stakeholders.
• Can: Practical experience with securing concrete AI use cases (LLM/RAG platforms, AI agents) in regulated environments and/or considering relevant AI and security regulations.

Project details

November 23, 2026

End: February 29, 2028

100%

Zurich and remote (only within Switzerland)

Inform yourself about your opportunities with Mr. Patrick Casutt or send us your CV directly by email to *@*.com. We look forward to hearing from you.

We refer to our data protection regulations (see on www.profilogic.com).

• Start: 2026-11
• Duration: 15 Monate
• Remote: hybrid
• Location: Zürich, Schweiz, CH
• Languages: German