Lakehouse platforms on Databricks and Microsoft Fabric.
Consulting and engineering for companies that want to turn their data into decisions they can rely on – scalable, maintainable and traceable.
Bronze, Silver, Gold – and everything that can go wrong in between.
A medallion architecture is quickly drawn. The difference lies in historisation, governance, infrastructure as code, tests and deployments – exactly where platforms fail in everyday use.
appendmergeviewsSeven things we are really good at.
Lakehouse architecture
Medallion design with Bronze, Silver and Gold layers, data contracts and governance via Unity Catalog.
- Target architecture and roadmap
- Naming conventions and layer model
- Permission concept
Databricks
Workspace setup, Lakeflow and Delta Live Tables, jobs, cluster strategy and performance tuning – on Azure, AWS or Google Cloud, most often on Azure.
- Workspace and catalog structure
- Pipeline frameworks
- Cost control
Microsoft Fabric
OneLake, Lakehouse and Warehouse, connection to Power BI and capacity planning.
- Fabric adoption
- Direct Lake and semantic models
- Capacity planning
Data engineering
Pipelines that keep running after handover: historisation, merge frameworks, orchestration, data quality, KPI models in the Gold layer.
- Historisation and surrogate keys
- Reusable merge utilities
- Data quality checks
Infrastructure as Code
We define Databricks workspaces, Unity Catalog, storage, networks and Fabric capacities as code – reproducible across DEV, QAS and PROD, traceable in Git and redeployable at any time.
- Terraform for Azure, AWS, Google Cloud and Databricks, Bicep for Azure
- Databricks Asset Bundles for jobs, pipelines and clusters
- Permissions, secrets and networking as code
- Drift detection and documented target states
CI/CD and DevOps
Git workflows and automated deployments across DEV, QAS and PROD – for code, pipelines and infrastructure alike.
- Branching and release strategy
- Automated tests and quality gates
- Azure DevOps and GitHub Actions
Migration and decision papers
Replacing existing data warehouses or switching platforms – with a paper that management understands.
- As-is analysis and scope
- Options with costs and risks
- Step-by-step migration
How we work with your team.
Project lead: architecture and engineering
We take the technical lead of your data project – architecture, infrastructure, standards, code reviews and implementation of the critical parts.
- For companies building or migrating a platform
- Project-based, usually over several months
- Remote, with regular on-site sessions
Technical advisory and review
You have your own team but need a partner who checks decisions, assesses architectures and helps with Databricks, Fabric or infrastructure questions.
- For companies with their own data team
- Hour budget or fixed days per month
- Available at short notice
A typical project.
Workshops with business and IT, review of the source systems, clarification of goals.
Target picture, layer model, governance, cost estimate and decision paper.
Infrastructure as code, pipelines and first data products – in two-week iterations with a demo.
Documentation, training for your team, optional ongoing support.
Frequently asked questions
Databricks or Microsoft Fabric – which one fits us?
That depends on team size, your existing Microsoft stack, data volume and the engineering depth you need. Fabric is often the faster entry for Power BI-oriented teams, Databricks the more robust choice for complex engineering workloads. We assess both for your case – also in combination.
Does it have to be Azure?
No. Databricks runs on Azure, AWS and Google Cloud, and we work on all three – most often on Azure, because Microsoft Fabric, Entra ID and Microsoft 365 come with it. The lakehouse architecture stays largely the same; the differences are in networking, identities and infrastructure as code, and we know those for all three providers.
Do you work remotely or on site?
Mostly remotely, with on-site sessions for workshops, kick-offs and key decisions. For clients in northern Germany a fixed day per week on site is also possible.
How does the engagement work contractually?
Project-based – on a time-and-materials basis or at a fixed price for clearly scoped packages such as a decision paper or a platform setup. For ongoing support we offer fixed monthly budgets.
Do you also support existing teams?
Yes. A large part of our projects is reinforcing existing data teams – as technical lead, in reviews or on individual topics such as historisation, performance, IaC or CI/CD.
And the “AI” part?
A good data platform is the prerequisite for any AI initiative. We build the foundation – clean, governed, current data – and connect models via Databricks and Fabric. For model development itself we work with specialised partners.
Intro call
Which question should your data platform answer?
Tell us about it. We will tell you openly whether Databricks, Fabric or something else is the right way.