ENNL
Book a call

Data platforms

Why a modern data platform pays off.

When data is spread over systems, reports take days. A governed platform aims to give you one trusted source, safe change and visible cost.

Book a callDatabricks health check

01 / The problem

The cost of data without a platform

These four patterns appear in almost every organisation that comes to us.

  • Data spread over systemsEach team keeps its own copy, so the same question gets different answers.
  • Slow, manual reportsNumbers are compiled by hand from several sources, so reporting takes days and depends on a few people.
  • Unclear ownershipNobody can say who owns a dataset, who may see it or where it came from.
  • Rising cloud costSpend grows without owners or limits, and nobody can explain the invoice.

02 / What changes

What a governed platform changes for you

A platform is not a product you buy. It is a set of decisions made once, in code, for every team.

Without a platform

  • Data spread over systems
  • Slow, manual reports
  • Unclear ownership
  • Rising cloud cost

With a governed platform

  • One governed source
  • Access per schema, with lineage
  • Safe change through code
  • Cost you can see and control
  1. Source systemsOn premises and Azure
  2. PipelinesBatch and streaming

  3. Governed by Unity Catalog

    1. LakehouseDelta Lake
    2. Access per schemaEntra ID groups
  4. Reports, AI and apps

What each one covers

  • One governed source. Shared data with named owners, so reports and AI use cases can start from the same numbers.
  • Access per schema, with lineage. Each team sees the schemas it needs and nothing more, and lineage shows where every number comes from.
  • Safe change through code. Development, test, acceptance and production run the same code, so a change can be reviewed and repeated.
  • Cost you can see and control. Owner tags, cluster policies and usage reporting are designed in from the start.
  • Databricks
  • Unity Catalog
  • Delta Lake
  • Azure
  • Terraform
  • Microsoft Entra ID

03 / Governance and security

Secure and private by design

These are our delivery standards, not options. They are what an audit asks about.

  • Private network pathsPrivate connections for every service, public network access switched off and Databricks running inside your private network.
  • IdentityAccess through Entra ID groups, and admin rights that are time-limited and need a stated reason.
  • SecretsA secure vault for secrets per environment and identities that need no stored passwords, so no long-lived keys sit in code or pipelines.
  • Data protectionClassified data, sensitive fields hidden or pseudonymised for people who do not need them, and EU regions for EU clients.

Governance for AI

One gateway for every model, agent and tool

  1. Teams, agents, coding toolsWhoever calls a model
  2. Unity AI GatewayPermissions, budgets, limits, policies, logging
  3. Models, MCP servers, toolsDatabricks-hosted and external
  • AccessUnity Catalog privileges decide who may use what.
  • Budgets and limitsSpend and request rates have a cap per use case.
  • PoliciesRules apply to requests and responses.
  • LoggingEvery call lands in Delta tables for audit.
  • RoutingTraffic splits across models, with failover.

How we set it up

  • The gateway and its policies are defined as code and deployed by pipeline, never clicked together.
  • One service per use case, each with its own budget and an owner who answers for it.
  • Inference tables with a retention period and access groups, so logs are useful and not a liability.
  • System-table dashboards for tokens, cost and latency per team.
  • Coding agents such as Claude Code are routed through the gateway, with repository rules and review gates on top.
  • Fallbacks are tested before go-live, by switching a destination off on purpose.
  • Smart Routing stays out of production while it is in Beta.

Generally available. Smart Routing is in Beta. Last verified: 10 November 2026. Unity AI Gateway documentation

  • Databricks
  • Unity Catalog
  • Unity AI Gateway
  • Delta Lake

Apps and analytics

OLTP, OLAP and Lakebase

Applications need a transactional database and analytics needs a lakehouse. We explain the difference, what Lakebase is and how we engineer one platform for both.

OLTP, OLAP and Lakebase in depth

ML and LLM operations

Models and LLM apps that stay in production

We build the path to production on Databricks: registry, pipelines, an evaluation gate, serving and monitoring, for classical models and LLM applications alike.

ML and LLM operations in depth

04 / Cost

Cost is a design decision

We do not promise a saving. We make spend visible and controllable, so you can decide.

  • Resources tagged by owner, so spend can be reported per team
  • Limits and auto-scaling that cap how much computing a workload can use
  • A review of cost and usage in the Databricks health check, ending in a prioritised fix list
  • Streaming and compute chosen for the latency you need, not the most you can buy
Cost viewIllustration
  • Nightly jobsTagged

    Every job has an owner and a cost centre.

  • Compute policyCapped

    Clusters are capped per team. Larger sizes need an approved request.

  • Streaming workloadReview

    Runs day and night but is only used in office hours. Proposal: schedule it.

  • Monthly reviewReported

    Spend per owner sent to each budget holder.

05 / Proof

Platforms in production

Outcomes are quoted from our case studies. Your result depends on your starting point.

All projects

06 / Start here

Two ways to begin

Already on Databricks or Azure data? Start with the health check. Building new? Start with the foundation.

Typical duration: 3 days

Databricks health check

We check cost, security with the Security Analysis Tool, governance with Unity Catalog and delivery with Asset Bundles and CI/CD, and give you a prioritised fix list.

What you get

  • Cost and usage review
  • Security and governance findings
  • Delivery and CI/CD review
  • Prioritised fix list with effort
Ask for a quote

Typical duration: 4 to 8 weeks

Azure data platform foundation

Landing zone, Databricks, Unity Catalog, CI/CD and Terraform modules, delivered secure by design.

What you get

  • Landing zone and network with private endpoints
  • Workspace, catalogue and access groups
  • CI/CD with build validation
  • First pipelines in production
Ask for a quote

Each starts with a conversation about your scope. Data platforms

Free check

Take the check

A score from 0 to 100 and a next step, in about three minutes.

All six checks

07 / FAQ

Questions about data platforms

Which platform do you build on?

Databricks on Azure. One platform for data engineering, analytics, machine learning and AI, governed in Unity Catalog, with Power BI and the rest of your Microsoft estate on top. Our references and credentials are all on Databricks.

We already have a platform. Is it worth a review?

Yes, that is what the Databricks health check is for: 3 days across cost, security with the Security Analysis Tool, governance with Unity Catalog and delivery with Asset Bundles and CI/CD.

How long does it take?

A foundation typically takes 4 to 8 weeks to the first pipelines in production. A health check takes 3 days.

Who owns the platform afterwards?

You do. Everything is code in your repositories, documented, with knowledge transfer to your team.

Next step

Is your platform ready for production?

Tell us how your Databricks and Azure platform runs today. We will say what to fix first, and whether a health check fits.

Discuss a health checkSend a messageor email info@nexops.nl directly

Currently accepting projects for Q4 2026