AI Architect - Databricks

Independent Consultant positionHybrid, 2-3 days per week in StockholmLong term programFull time - 40 hours per week We are supporting a major enterprise modernisation programme and are looking for an AI Architect with strong Databricks expertise to help define and deliver the next phase of the organisation’s Data & AI platform. This role sits at the intersection of AI engineering, Lakehouse architecture, governance, and cloud platform strategy, and is well‑suited to someone who enjoys shaping standards, guiding engineering teams, and building scalable foundations for advanced analytics and LLM workloads. ResponsibilitiesLead the architectural design of the Databricks AI Platform, covering data, ML, LLMOps, governance, and production deployment.Define and evolve Lakehouse architecture using Delta Lake, medallion patterns, and scalable data engineering frameworks.Architect ML & LLM pipelines using MLflow, Feature Store, Model Serving, and Databricks Workflows.Implement and standardise Unity Catalog governance, including security models, lineage, access patterns, and workspace strategy.Partner with data scientists and ML engineers to operationalise models, automate retraining, and ensure reliable productionisation.Establish DevOps and CI/CD practices for Databricks, enabling automated deployment, testing, and environment consistency.Work closely with cloud platform teams (Azure preferred) to ensure alignment with enterprise architecture, compliance, and cost‑efficiency.Provide architectural guidance, run workshops, and support decision‑making across multiple stakeholder groups. Experience RequiredStrong experience architecting solutions on Databricks (Azure, AWS, or GCP).Deep knowledge of Delta Lake, Unity Catalog, cluster policies, and Lakehouse governance.Hands‑on expertise with PySpark, SQL, and modern data engineering patterns.Experience designing ML/LLM pipelines, including MLflow, Feature Store, vector search, embeddings, or RAG patterns.Solid understanding of cloud architecture (Azure preferred) and enterprise security.Comfortable operating in large‑scale programmes with multiple teams and senior stakeholders.