Databricks AI Architect

About the role

You will own the AI and data architecture for enterprise clients. You will design Lakehouse platforms on Databricks. You will take GenAI and ML use cases from idea to production. You will be the senior technical voice in front of the client.

What you will do

• Design end-to-end Lakehouse architectures on Databricks: ingestion, Delta Lake, medallion layers, serving.
• Architect GenAI solutions: RAG, vector search, agents, model serving, and evaluation.
• Lead ML architecture using MLflow, Feature Store, and Model Serving.
• Set up governance with Unity Catalog: access control, lineage, data classification, audit.
• Define security and data residency patterns that meet local regulations and client policies.
• Build CI/CD and MLOps standards using Databricks Asset Bundles, Git, and Terraform.
• Run architecture workshops, solution reviews, and design sign-offs with client stakeholders.
• Size platforms and optimise cost and performance (clusters, SQL warehouses, Photon, serverless).
• Guide and review the work of data and AI engineers. Raise the bar on code and design quality.
• Support pre-sales: solution designs, estimates, and technical sections of proposals.

What you must have

• Hands-on Databricks experience on production projects (mandatory).
• Senior background in data engineering, data architecture, or ML/AI.
• Proven delivery of at least two enterprise Lakehouse or AI programs as lead architect.
• Strong Spark, PySpark, and SQL. Solid Python.
• Deep knowledge of Delta Lake, Unity Catalog, MLflow, and Databricks workflows.
• Hands-on GenAI work: LLMs, RAG, vector search, prompt and model evaluation.
• Experience on at least one major cloud: Azure, AWS, or GCP.
• Clear, confident communication with senior business and IT stakeholders.

Nice to have

• Databricks certifications: Data Engineer Professional, ML Professional, or GenAI Engineer Associate.
• Experience in regulated sectors: government, banking, energy, or telecom.
• Knowledge of data protection and data sovereignty requirements.
• Background migrating from legacy warehouses (Teradata, Oracle, Hadoop) to Databricks.

Success in the first 6 months

• Target architecture signed off by the client.
• First AI use case live in production.
• Governance and CI/CD standards adopted by the delivery team.