AI engineer in vulnerability management

We are building the next generation of intelligence in our complex multi-service environment that connects multiple systems, data sources, and workflows at scale achieving real-time intelligence for product transparency, vulnerability operations. As Staff AI Engineer, you own the AI functions end to end including what they must do, how they are evaluated and accepted, how they integrate, and which risks they introduce. In this role, you set technical direction, make evidence-based go/no-go decisions, and influence several teams, without direct reports, line management, or personnel responsibility.

• Write the use-case specification for each AI function, covering inputs, expected behavior, success criteria, failure modes, and human decision points.
• Agree with capability owners what is automated, assisted, or requires approval.
• Design and run the evaluation approach: golden datasets from our triage and remediation history, correctness and safety metrics, offline and online evaluation, and drift monitoring in production.
• Hold the acceptance gate at each stage (build or buy, pilot, limited and general availability) and make go/no-go decisions on evidence.
• Escalate unresolved disagreements through program governance.
• Define where models sit, how data flows, where humans stay in the loop, and what evidence is retained for audit.
• Review every integration design that touches AI functions.
• Own the AI threat model, define and test controls against prompt injection, data leakage, unsafe actions, and model supply-chain risk, and lead the response when an AI function misbehaves.
• Act as the technical counterpart to the AI technology provider on quality requirements, evaluation obligations, model-change notification, and defect handling.
• Own the standard for AI use in the team’s own engineering, including approved tools, data handling, review of AI-generated code, provenance, licenses, and secrets.
• Guide and mentor the engineers working on AI functions and build evaluation skills across teams.
• Build datasets, run evaluations, and review code wherever that resolves uncertainty.

• 10+ years in software engineering, including several years shipping and operating LLM-based or other ML-based features in production.
• Correctness and safety metrics, offline and online evaluation, drift monitoring, and launch decisions based on that evidence.
• Designing and reviewing service integrations, data flows, human-in-the-loop points, and evidence retention.
• Threat modeling AI systems and controls against prompt injection, data exfiltration, and unsafe tool use.
• Working knowledge of advisories, CVEs, SBOMs, and remediation.
• Writing acceptance criteria that engineers and suppliers are held to.
• Production-quality code in Python, Go, Java, or Rust.
• Leading technical work across teams without formal authority, mentoring engineers, and saying no on evidence when that is unwelcome.
• Explaining risk to executives and constraints to developers, with excellent written and spoken English.
• Будет плюсом: Evaluating code-generating models (patch correctness, test-based validation, regression risk); Acting as the technical counterpart to an AI vendor or internal provider; Software composition analysis or secure development tooling; Familiarity with the EU Cyber Resilience Act and the EU AI Act; Building an evaluation capability from scratch

• Hybrid work model that supports flexibility and a sustainable work–life balance.
• An environment where everyone can bring their whole self to work and feel a sense of belonging.
• A workplace that values curiosity and guarantees continuous learning, with development opportunities for both personal and professional growth.
• Share matching programs to become a shareholder of Siemens AG.
• A wide range of flexible benefits for you and your family to enjoy