AI Engineer - AI SDLC Tooling & Guardrails

BME - Bolsas y Mercados Españoles - impulsa la transformación de los mercados financieros y pertenece a SIX, el tercer grupo bursátil más grande de Europa.

Lo que nos hace diferente, nos impulsa a avanzar: entre raíces locales y relevancia global, somos una mezcla única de tradición y futuro, de cimientos y crecimiento. Apreciamos el valor de mentes brillantes y las inspiramos a crecer con sus ideas. Da forma al futuro de las finanzas con nosotros.

AI Engineer - AI SDLC Tooling & Guardrails

Warsaw, Madrid|Working from home up to 40%| Referencia8247

Agentic AI is changing how software gets built - and at SIX, it will touch the work of more than thousand engineers. We're building a new team at the center of this shift, responsible for shaping how AI-driven software engineering works across the company. As part of this team, you will build the guardrails that make agentic development trustworthy at scale: the test automation, coverage enforcement, and independent review layers that let AI agents write and ship code without compromising quality.

What You Will Do

• Build AI SDLC tooling as a platform capability, not a project. Its pillars are evaluation of agents, skills and prompts; AI code review in CI; and test automation: agent-driven test generation and maintenance, coverage enforcement on changed lines, flaky-test detection and self-healing, mutation testing, and test-suite health (redundancy, gaps, execution time) across unit, integration, and end-to-end layers.
• Build and operate the AI agent harness. The sandboxed execution environment, tool/permission scoping, and CI/CD integration that lets agents propose, test, and merge code, instrumented with decision logs and action provenance so every agentic change is traceable and reproducible.
• Own the independent review layer for AI-generated code and tests, including AI code review measured for missed defects and false positives: correctness, coverage, and quality gates in CI/CD that verify what an agent produces using tooling the agent doesn’t control, so speed never comes at the cost of test quality or code health.
• Define and evolve the quality and tooling strategy for an AI-driven SDLC: what gets tested where, how test data is generated and managed, how test quality is measured (not just pass/fail, but whether the tests catch regressions), and how the SDLC standards defined by the Enterprise SDLC track are enforced in tooling.
• Treat AI SDLC tooling and guardrails as a product, not a gate. Instrument for adoption, false-positive rate, and developer trust, keeping the safe, well-tested path the fastest path.

What You Bring

• Hands-on familiarity with agentic AI architectures: agent orchestration, tool-use/sandboxing, and harness design, not just conceptual awareness of “AI in the SDLC.”
• 6+ years in software engineering with deep, hands-on expertise in developer tooling or platform engineering. Test automation experience is a strong plus, not a prerequisite. Ideally someone with a track record of building tooling or developer infrastructure that other engineering teams adopted, not just writing test suites.
• Experience evaluating LLM or agent systems (eval sets, error analysis, LLM-as-a-judge) is a strong plus.
• Working knowledge of test methodology and code quality: test pyramids, coverage strategy, mutation testing, flaky-test remediation, and what makes a test suite trustworthy versus just green.
• Platform-product mindset. You build reusable capabilities and pave roads for other engineers, and you measure and strive for adoption and impact across SIX’s business units.
• Working knowledge of governance and controls in a regulated or high-assurance setting. Financial market infrastructure experience is a plus; experience in an adjacent regulated industry counts too.

Si tienes alguna duda, consulta nuestra página de preguntas frecuentes (FAQ) o contacta con Yuliya Stoyko+34 917095993.

Para esta vacante sólo aceptamos solicitudes directas.

La diversidad es importante para nosotros. Por lo tanto, estamos deseando recibir solicitudes independientemente de cualquier contexto personal.