Skip to content
woodfrog

·

Data Foundations
Altysys, delivering for retail and FMCG clientsCustomer story, named. Results to be published with Altysys

The Databricks team behind Altysys.

Altysys wins and runs data projects for retail and FMCG companies. When a client needs Databricks, woodfrog builds it: the pipelines and lakehouse, the machine learning, the dashboards and the governance. Altysys keeps the client. woodfrog delivers the engineering.

One delivery team, two companiesSchematic
AltysysClient relationshipDomain knowledgeDelivery oversight
woodfrog on DatabricksPipelines and lakehousePredictive maintenance MLDashboards and BIMigration and governance
For Altysys's clientsRetail and FMCG
4Databricks capabilities delivered
Retail · FMCGSectors of Altysys's clients
3 peopleA senior woodfrog pod per engagement
1 teamFrom scoping to support, no handover
WhoAltysys, a technology services company with retail and FMCG clients
EngagementAn ongoing delivery partnership. Altysys owns the client; woodfrog delivers on Databricks.
BuiltLakehouses and pipelines, predictive maintenance models, dashboards, and migration and governance on Databricks
Before and after

A partner that can say yes to Databricks work.

Before
  • Client asks for Databricks, and the team has to hire or hand off
  • Each project starts from a blank page
  • Data engineering, machine learning and BI split across vendors
  • Governance added late, after the client asks for it
After
  • A senior woodfrog pod ready when a client needs Databricks
  • Our lakehouse blueprint, data contracts and quality checks on every project
  • Pipelines, models and dashboards built by one team
  • Unity Catalog access and lineage from the first table
The problem

Clients want Databricks. Building a bench for it is slow and costly.

Retail and FMCG clients increasingly ask for Databricks: a lakehouse for sales and stock, models that predict failures, dashboards the business trusts. For a services company, keeping certified Databricks engineers on the bench for work that arrives in waves is expensive, and turning work away loses the client.

What was broken
  • Databricks work arrives in waves, so a full-time bench sits idle between projects.
  • Each client needs pipelines, machine learning and BI, which rarely sit in one hire.
  • Governance and cost control are easy to skip under deadline, and hard to add later.
  • Turning down a Databricks request sends the client to another vendor.
You might be here if
  • Your clients ask for Databricks and you do not have the team.
  • You need machine learning and BI as well as data engineering.
  • Your name is on the delivery, so quality cannot slip.
  • You want to keep the client relationship yourself.

Why it mattered. The partner's name is on the work. It needs a delivery team it can trust with its clients.

The approach

Altysys keeps the client. woodfrog builds on Databricks.

Altysys owns the relationship, the domain knowledge and delivery oversight. woodfrog joins at scoping and builds the Databricks work end to end: data in, models and dashboards out, governed throughout. Altysys reviews every release before it reaches the client.

Altysys
Wins and runs the client
Brings domain knowledge
Reviews every release
woodfrog pod
Pipelines and lakehouseDelta Lake, tested layers
ML and dashboardsModels in production, BI on top
GovernanceUnity Catalog from table one
The client gets
One lakehouse for sales and operations
Models and dashboards in use
Databricks, delivered under Altysys
Deliberately absentA handover between a data vendor, an ML vendor and a BI vendor
The same pod stays from scoping to support, so the people who built it answer the questions.schematic
The solution

Four kinds of Databricks work, one team.

What was built
  • Pipelines and lakehouses: sales, stock and operations data landed in Delta Lake in bronze, silver and gold layers, scheduled and tested.
  • Predictive maintenance: models on equipment sensor data that estimate remaining useful life and flag anomalies, with a monitoring dashboard and alerts on model performance.
  • Dashboards and BI: governed metrics and dashboards on top of the lakehouse for business teams.
  • Migration and governance: moving client data onto Databricks, with Unity Catalog access, masking and lineage.
Each model is deployed behind an API with a simple interface, and its behaviour is logged so performance can be tracked and signed off over time.
The impact

What the partnership delivers.

Delivered on Databricks for Altysys clients
LakehousePipelines in tested layers, on Delta Lake
Predictive maintenanceRemaining useful life and anomaly detection
DashboardsGoverned metrics for business teams
GovernanceMigration and Unity Catalog access
What we track with AltysysFigures to be published with Altysys's approval
Client projects delivered on DatabricksCounted
Model accuracy and alerts acted onPer model, monitored
Dashboards in daily usePer client
Tables governed in Unity CatalogShare of all tables
Need a Databricks delivery team behind your brand?A short call, or record a message. A person replies within one business day.
Contact usCertified Services Partner, Claude Partner Network