Security Architect, Agentic AI

About Zscaler Zscaler accelerates digital transformation to ensure our customers can be more agile, efficient, resilient, and secure. As an AI-forward enterprise, we are constantly pushing the envelope, leveraging the world’s largest security data lake to power our cloud-native Zero Trust Exchange platform. This innovation protects our customers from cyberattacks and data loss by securely connecting users, devices, and applications in any location. Here, impact in your role matters more than title and trust is built on results. We say, impact over activity. We seek innovators who actively use AI to amplify their impact and who thrive in an environment where we leverage intelligent systems to stay ahead of evolving threats. We believe in transparency and value constructive, honest debate—we’re focused on getting to the best ideas, faster. We build high-performing teams that can make an impact quickly and with high quality. To do this, we are building a culture of execution centered on customer obsession, collaboration, ownership, and accountability. We value high-impact, high-accountability with a sense of urgency where you’re enabled to do your best work and embrace your potential. If you’re driven by purpose, thrive on solving complex challenges, and want to be part of the team that’s helping to secure the AI age, we invite you to bring your talents to Zscaler and help shape the future of cybersecurity. Role We are looking for an Agentic AI - Architect/Security Engineer to join our team. This is a fully remote U.S. role, reporting to the VP, Information Security Engineering in the Exposure Management & Security Operations department. We’re looking for a hands-on Security Architect/Engineer to build and ship security capabilities for agentic AI systems—LLM apps that plan, use tools, manage memory, and leverage enterprise data. This is an implementation-focused role where you will deliver production code, secure building blocks, and reference implementations for both IT and Product teams. What you’ll do (Role Expectations) - Build secure agent runtimes, libraries, and reference implementations while implementing core agent patterns like planner/executor, tool routing, and RAG boundaries - Build and secure MCP servers, clients, tool registries, and connector patterns with robust authentication, authorization, and audit logging - Enforce secure-by-default controls including schema validation, tool allowlists, redaction, and policy checks - Threat model and test agent workflows for prompt injection and data exfiltration to build repeatable security evaluations Who You Are (Success Profile) - You thrive in ambiguity. You're comfortable building the path as you walk it. You thrive in a dynamic environment, seeing ambiguity not as a hindrance, but as the raw material to build something meaningful. - You act like an owner. Your passion for the mission fuels your bias for action. You operate with integrity because you genuinely care about the outcome. True ownership involves leveraging dynamic range: the ability to navigate seamlessly between high-level strategy and hands-on execution. - You are a problem-solver. You love running towards the challenges because you are laser-focused on finding the solution, knowing that solving the hard problems delivers the biggest impact. - You are a high-trust collaborator. You are ambitious for the team, not just yourself. You embrace our challenge culture by giving and receiving ongoing feedback—knowing that candor delivered with clarity and respect is the truest form of teamwork and the fastest way to earn trust. - You are a learner. You have a true growth mindset and are obsessed with your own development, actively seeking feedback to become a better partner and a stronger teammate. You love what you do and you do it with purpose. What We’re Looking for (Minimum Qualifications) - Foundational understanding of AI/ML technologies and experience leveraging, securing, or positioning AI-driven solutions to optimize outcomes within your functional domain - Extensive experience designing and implementing agent runtimes including planning, tool calling, and memory/state controls - 8+ years of experience shipping production security engineering solutions with strong coding skills in Python, TypeScript, or Node - Practical experience integrating frontier model APIs with structured outputs, streaming, and safety controls - Familiarity with agent frameworks such as LangGraph, LangChain, or Semantic Kernel securely paired with solid cloud and Kubernetes security engineering experience What Will Make You Stand Out (Preferred Qualifications) - Experience designing advanced guardrail architectures or real-time alignment enforcement mechanisms for autonomous multi-agent orchestration - Familiarity with OWASP LLM Top 10, MITRE ATLAS, or the NIST AI RMF, and experience building red-team style testing or evaluation harnesses for LLM applications and ag

Foundational understanding of AI/ML technologies and experience leveraging, securing, or positioning AI-driven solutions to optimize outcomes within your functional domain; Extensive experience designing and implementing agent runtimes including planning, tool calling, and memory/state controls; 8+ years of experience shipping production security engineering solutions with strong coding skills in Python, TypeScript, or Node; Practical experience integrating frontier model APIs with structured outputs, streaming, and safety controls; Familiarity with agent frameworks such as LangGraph, LangChain, or Semantic Kernel securely paired with solid cloud and Kubernetes security engineering experience.

Build secure agent runtimes, libraries, and reference implementations while implementing core agent patterns like planner/executor, tool routing, and RAG boundaries; Build and secure MCP servers, clients, tool registries, and connector patterns with robust authentication, authorization, and audit logging; Enforce secure-by-default controls including schema validation, tool allowlists, redaction, and policy checks; Threat model and test agent workflows for prompt injection and data exfiltration to build repeatable security evaluations.