Associate Director, AI Security Architecture

Location: Israel

Job Description Summary

Internal Job Title: Assoc. Dir. DDIT ISC Security Architecture AI

#LI-Hybrid

Location: Israel

Relocation Support: This role is based in Israel. Novartis is unable to offer relocation support: please only apply if accessible.

Help shape how AI can be built, deployed, and scaled securely. In this role, you’ll influence the security architecture behind next-Generation AI solutions, defining standards and patterns that protect AI models, data, pipelines, and agentic systems throughout their lifecycle. You’ll work at the intersection of AI, cybersecurity, cloud engineering, and architecture, partnering with technical teams to address emerging threats, strengthen secure AI development, and turn evolving security challenges into practical, scalable solutions.

Key Responsibilities

• Lead AI architectural activities and develop secure solution architectures across business and technology domains.
• Define AI security standards and best practices for protecting AI pipelines, datasets, models, and infrastructure.
• Develop logical architecture designs and strategies to secure the organization’s AI domain and infrastructure.
• Establish and evolve AI architecture governance frameworks, including methods, practices, standards, and reusable architecture patterns.
• Assess the AI landscape, identifying and mitigating security risks across model training, deployment, and operations.
• Evaluate AI security technologies and integrate appropriate security controls across AI platforms and Large Language Models.
• Secure AI and agentic systems against model poisoning, adversarial inputs, prompt injection, jailbreaking, and emerging threats.
• Design Model Context Protocol servers, AI agent workflows, secure deployment architectures, and AI gateway integrations.
• Build Python services, developer tools, integrations, and AI security harnesses supporting secure software development.
• Collaborate with data scientists, engineers, and DevOps teams to embed security throughout the AI development lifecycle.