Databricks Data Architect – Healthcare & Life Sciences

About the Role

We’re hiring on behalf of a specialist Databricks consultancy — an inner-circle partner with 150+ Databricks projects delivered and direct access to product roadmaps and private previews.

This is architecture, not model-building. You’ll own end-to-end AI lifecycles for global healthcare and life sciences clients: computer vision pipelines on large-scale medical imaging, running in production on the Databricks Lakehouse, under HIPAA, GDPR and HITRUST constraints that shape the design from day one.

It sits at an unusual intersection — deep learning depth, data engineering rigour, and a consultant’s mindset. You’ll advise clients on clinical validation, build-vs-buy decisions and long-term AI strategy, then go build the thing. If you’d rather stay in a notebook and hand off to someone else for production, this isn’t your role.

What You’ll Do

Computer Vision & Imaging Pipelines

• Design and implement computer vision pipelines in PyTorch or TensorFlow on Databricks Runtime for ML, processing large-scale medical imaging datasets (DICOM, NIfTI).
• Build efficient data loaders and preprocessing for high-dimensional medical data, parallelising transformation and feature extraction with Apache Spark.

Production MLOps

• Architect production MLOps with MLflow — experiment tracking, model versioning, and the handover from research into production Model Serving.
• Standardise delivery with CI/CD for ML (Git integration, Databricks Asset Bundles) and Feature Stores that keep training and real-time inference consistent.

Governance & Compliance

• Enforce data lineage and security through Unity Catalog, with training and inference meeting healthcare regulatory standards (HIPAA / GDPR / HITRUST).

Technical Leadership

• Advise clients on build-vs-buy for medical AI tooling, model interpretability, and bias mitigation in clinical settings.

Requirements

• 5+ years in Machine Learning or Data Science, including 3+ years hands-on deploying models inside the Databricks ecosystem. The Databricks depth is non-negotiable.
• Proven work with medical imaging formats — DICOM, NIfTI, WSI.
• Advanced Python (PyData stack) and SQL, with PySpark for large-scale data manipulation.
• Production experience on Azure (Azure ML, ADLS Gen2) or AWS (SageMaker, S3), including GPU instance management and cost control.
• MLflow across the full lifecycle, and Delta Lake for managing unstructured imaging metadata.
• Strong grounding in deep learning architectures (CNNs, Transformers, SegNet) and the evaluation metrics that matter in medical diagnostics.
• English at C1+ — you’ll work directly with clients and translate algorithmic detail into business decisions for non-technical stakeholders.

Nice to Have

• Specialised imaging libraries: MONAI, SimpleITK, OpenCV.
• Databricks Model Serving or TorchServe in production.
• Unity Catalog for lineage and access control at scale.
• Hands-on delivery under HIPAA or HITRUST, not just awareness of them.
• Databricks Machine Learning Professional certification, or a specialised cloud AI certification.

What’s On Offer

• 100% Remote — full flexibility on where you work from.
• B2B contract — cooperation model is B2B only.
• High-end hardware — Apple MacBook Air M4 15” provided.
• Certification fully covered — all Databricks technical certifications.
• Career growth — a track to the top tiers of Databricks expertise (Champion-level programmes), shaped around your career direction.
• Expert mentorship — from Databricks MVPs, with collaboration alongside core product teams.
• Visibility — conference speaking and thought-leadership opportunities if personal branding matters to you.
• Referral bonus — for bringing other senior engineers into the team.

Rates are set per profile. Tell us your expectation on the first call and we’ll tell you straight away whether it’s in range — no wasted rounds.

Hiring Process

1. Introductory call (20 min) — background, expectations, rate range, mutual fit.
2. Technical interview (60 min) — computer vision, Databricks architecture, and production MLOps with the engineering team.
3. Client interview — with the healthcare client, where the project requires it.
4. Decision & offer — fast feedback either way, no radio silence.

We’re former software engineers, so the screening call is a technical conversation, not a keyword check. We’ll tell you what the project actually looks like — including the parts that are hard.