AI System Architect

Tufin · telaviv

Description Tufin is standing up a governed, enterprise-scale AI program that spans ChatGPT, Claude, Workato eMCP, and a growing ecosystem of third-party AI applications. The AI System Architect is the most senior technical role in this program — the person who defines the architecture, enforces the governance model, and owns the integration surface that every AI agent in the company operates through.  This role sits inside Enterprise Technology, reporting directly to the Head of Enterprise Technology. That placement is intentional. The AI System Architect is not a researcher, a prompt engineer, or a standalone AI strategist — they are an enterprise systems leader who happens to be building at the frontier of agentic AI. They own the AI integration strategy across Tufin's core platforms (Salesforce, NetSuite, Workato, HiBob, Jira), the MCP governance model, the persona-scoped token design, and the integration patterns that connect AI capabilities to those systems without creating point-to-point dependency risk.  You will manage the AI Platform Engineer(s), set the technical standards for the AI Power User group's citizen development program, and serve as the connective tissue between business leadership, platform owners, and development teams. You will shape the multi-year AI architecture roadmap while also rolling up your sleeves to conduct architecture reviews, resolve blockers, and move use cases from concept to production. This is a role for someone who can think big and execute — and who understands that in an enterprise context, the quality of your governance is inseparable from the quality of your architecture.  What You'll Own   Strategy & Architecture   Define and own the enterprise AI integration strategy — identifying opportunities to embed intelligent automation, agentic workflows, predictive analytics, and generative AI capabilities across Tufin's core platforms  Develop and maintain reference architectures, design patterns, and the AI architecture decision log that governs how AI models connect to enterprise systems and what they are permitted to do  Consult on enterprise system architecture and implement best practices for the Enterprise Business Systems team to leverage in their day-to-day execution.   Lead Proof-of-Concept initiatives for new AI tools and platform-native AI features, evaluating them against build-vs-buy criteria before recommending adoption  Partner with business stakeholders to translate operational pain points into AI use cases with clear ROI framing and sequencing criteria  Contribute to Tufin's enterprise data strategy, ensuring AI initiatives are supported by clean, accessible, and well-governed data pipelines  Integration Architecture & Delivery   Design and own the Workato eMCP layer — the MCP governance model, persona-scoped token framework, workspace isolation strategy, and the single sanctioned action surface through which all AI agents write back to enterprise systems  Define integration patterns and standards for AI model connectivity (Claude, ChatGPT) to Salesforce, NetSuite, HiBob, and Jira — specifying what agents can read, what they can write, through which surfaces, and with what confirmation and audit requirements  Design and oversee API strategies, event-driven architectures, and middleware patterns that support scalable AI feature delivery — including agentic workflows, intelligent data transformation, anomaly detection, and natural language interfaces layered onto ERP and CRM data  Collaborate with Engineering during build phases, conducting architecture reviews, providing hands-on guidance, and resolving complex technical blockers  Define non-functional requirements — latency, security, auditability, model drift monitoring — for AI components embedded in mission-critical business processes  Establish MLOps and LLMOps practices appropriate for Tufin's enterprise environment: model versioning, observability, and rollback procedures for production AI workloads  Governance & Risk   Translate Tufin's AI governance framework into enforceable runtime controls: confirmation gates, role-scoped permissions, audit trails, and rate limiting across all production agents  Own the AI intake process — the structured gate through which new AI use cases, agent deployments, and integration requests are reviewed, approved, and sequenced  Lead AI impact assessments for enterprise use cases, accounting for data privacy, regulatory compliance (GDPR, SOC 2, and applicable industry mandates), and responsible AI principles  Partner with Tufin's Security and Compliance teams and AI Governance Committee to define guardrails for agents operating with write access to critical systems — including human-in-the-loop checkpoints and audit trail requirem

Apply »