AI Platform Engineer; Mosaic

AI Platform Engineer (Mosaic) page is loaded## AI Platform Engineer (Mosaic) locations:
Madrid time type:
Full time posted on:
Opublikowano dzisiajjob requisition :
REQ
2702813
** NN Digital Hub
** is a subsidiary company of
** Nationale Nederlanden Group
** located in Madrid, Spain. We deliver IT services and solutions for the different international Business Units from Nationale Nederlanden Group.

Our objective is to create, deliver and maintain insurance product management solutions, as service propositions to enable any agent (linked or direct), machine or company to create and sell insurance products in Europe in an automated way and real time.

We are looking for an experienced
** AI Platform Engineer
** to join our team.

Your role is to build and productize AI capabilities — from RAG pipelines to model serving — and deliver them as scalable, governed, and developer-friendly services.##
** Your impact at NN:
What you´ll do:**- You will turn Mosaic into a consumable platform product: a governed, observable, and reliable experience for building and running multiple AI workload styles (agentic, batch scoring, streaming/micro-batch, real-time APIs, scheduled/directed workflows) through a single, consistent developer experience.

- You will encode governance boundaries (Unity Catalog + Collibra) into developer workflows, provide RAG-ready components (Vector Search + MLflow evaluation), and standardize connectivity via protocols (MCP, A2A) and mediation layers, so teams can deliver fast without compromising security or cost.
* Design and operate multi-tenant Mosaic platform baselines (work spaces, policies, secrets, networking) with Azure/AWS parity and serverless vs K8s topologies.
* Golden paths & self-service:
Ship repo/pipeline/environment automation, SDK/CLI/templates covering agentic apps, classical ML services, and directed workflows; enable citizen-builder flows with quotas, scoped access, and pre-approved components.
* Provide RAG & evaluation reference implementations:
Vector Search indexing/retrieval, MLflow offline/online evaluation suites, dashboards for quality/latency/drift/cost.
* Embed governance-by-design:
Unity Catalog policies/lineage, integration with Collibra workflows and marketplace metadata; ensure approval trails are automatic.
* Enable protocol & integration patterns: implement MCP/A2A connectors, NN Mediation Layer interfaces, and downstream integration runbooks with SLOs and incident playbooks.

As part of our Employee Experience, we offer you a range of competitive benefits that are available to you in order to improve the physical, mental, and professional wellbeing of our workers. Among the benefits are:
* We work with a hybrid model
* Starting financial aid so you can equip your work space at home
* Allowance and telework subventions
* Flexible working hours and 2 months of intensive working hours in summer so you can fully take advantage of your time
* We hire a life insurance and a pension plan for all our employees
* We establish objectives bonus as a performance reward
* You can get to the office as you please. We have a free parking lot for cars, motorbikes, electric cars with chargers, and bikes.
* Get to know our flexible retribution facilities such as transport card, nursery checks, health insurance with Sanitas, trainings…
* We care about people. That is why we are involved in society by facilitating volunteer actions and time to all our employees.
* We care about your wellbeing. We have a Wellness Program available.
* We are digital and we love technology. Also, every team works under agile methodology.
* 6-8+ years software/platform engineering (Python) with production ML/AI platforms; hands-on Databricks is preferred (MLflow, Unity Catalog) and familiarity with agentic and classical ML delivery patterns
* Designed and operated multi-tenant platforms;
IaC; CI/CD; identity, secrets, network isolation, and policy-as-code.
* Experience across multiple workload types: real-time APIs, batch/directed workflows, and streaming/micro-batch.
* Hands-on with RAG systems (chunking/indexing/retrieval).
* Built evaluation/monitoring capabilities (hooks, metrics, alerting) and enforced gates in…