AI Engineer

• Engineering
• Tel Aviv
• Senior
• Full-time

Description

What we’re looking for:

We're looking for an AI Engineer to take a central role in shaping and evolving the intelligence layer of Linx, an AI-native identity platform. You’ll help build the AI brain behind our assistant, Autopilot, and the identity-specific agents that monitor, investigate, govern, and remediate identity risk for some of the world’s largest enterprises.

This is a hands-on role for someone who drives technical direction, raises engineering standards, and brings clarity to complex systems. At Linx, complexity is the baseline: you’ll work on a graph-based Identity Fabric that unifies human, non-human, and agentic identities across enterprise environments, where a wrong model output isn’t a demo bug, but a security risk.

You’ll work end-to-end, from research and design through production, partnering with backend, security research, product, and data teams to build scalable, reliable, and trustworthy AI systems.

What You'll Do

• Lead the design and implementation of AI-powered systems end-to-end, across our AI assistant, Autopilot (autonomous agents for real-time identity risk detection and remediation), and future identity-specific agents.
• Integrate LLMs into security-critical workflows, beyond prompting into tool use, structured outputs, function calling, retrieval strategies, model routing, and orchestration.
• Own the full lifecycle, from problem definition and data design to prompting, evaluation, and production, while defining quality and building the systems to measure it.
• Engineer for trust and control - guardrails, validation layers, and human-in-the-loop mechanisms that enable safe autonomy.
• Evolve our AI-native data layer (Identity Graph, DBs, analytics, APIs), recognizing that AI quality is fundamentally a data problem.
• Drive architectural decisions balancing scalability, latency, and cost across large, multi-tenant systems.
• Collaborate closely with backend, product, security research, and data teams across identity, automation, and governance domains.
• Establish best practices for building and deploying LLM systems: testing, observability, and reliable agent behavior at scale.
• Continuously experiment and refine: ship fast, measure with real signals, and improve across prompts, tools, retrieval, and data.

Requirements

What You'll Bring

• Extensive experience in backend, AI, or data-intensive systems, with a track record of building production systems used by real customers.
• Hands-on experience with LLMs in production, including prompting, tool use, function calling, orchestration, and evaluation, with a strong understanding of real-world failure modes.
• Strong ability to design and build end-to-end systems, from data pipelines and retrieval to deployment, monitoring, and iteration.
• Deep understanding of data design for AI systems, including schemas, semantics, and workflow structuring for reliable outputs.
• Strong coding skills in Python (or similar) and experience working in modern backend systems at scale, with awareness of latency, cost, and multi-tenancy.
• Proven ability to drive architectural decisions and technical direction in ambiguous, high-impact areas.
• Pragmatic, product-oriented mindset, focused on solving real customer problems with the right tools.
• Strong communication skills, with the ability to explain trade-offs and align across teams.
• Passion for AI and a builder mindset, with a focus on ownership, speed, and impact.

Nice to Have:

• Experience building Text-to-SQL, Text-to-Query, or NL-to-data systems over complex or graph-based data.
• Experience building agentic systems in production, including tool routing, long-running workflows, and failure handling.
• Hands-on experience with RAG, hybrid retrieval, semantic search, and learning from user interactions.
• Exposure to MCP, agent-facing APIs, or data access layer design for AI systems.
• Background in identity or security domains (IAM, IGA, PAM, etc.).
• Experience designing LLM evaluation frameworks, including benchmarks, metrics, and human-in-the-loop review.