One platform, three entities, zero data loss
DevOps engineer who split an enterprise data platform across 3 newly-spun-off companies. Kubernetes, Terraform, Redshift, zero data loss, hard regulatory deadline.
Who they are
A European tire manufacturer, enterprise-scale, multi-billion revenue, going through a corporate restructure that split the business into 3 independent legal entities. Each entity needed its own production infrastructure, data pipelines, and SAP integration – fully separated, fully working, no shared dependencies post-split.
What was on the table
Splitting a working infrastructure into three working infrastructures is one of those projects that looks like copy-paste and ends up being archaeology.
The custom data layer was the breaking point. Built over years by an internal team, it had implicit dependencies on Redshift configurations, on specific SAP endpoint behavior, on data flowing in a single shape. Once you start copying it across three environments, none of that survives. Things break in silence: data flows look correct, but values drift. Reports look fine, but the numbers behind them are wrong.
The client had a hard regulatory deadline for the split. Failing it wasn’t an option – it had legal and contractual consequences. They needed someone senior enough to own this domain, not just “execute tickets.”
Why they chose UASoftDev
This wasn’t a “send us a CV” engagement. We proposed a Dedicated Team model – one senior DevOps engineer leading the split, with Alexander (CTO) involved in initial architecture review and weekly checkpoints.
Our engineer had prior experience with both Terraform-managed AWS at scale and custom Redshift data layers – the specific combination needed for this split. First architectural call within 60 hours of brief.
What the engineer delivered
- Mapped the existing infrastructure to identify all implicit dependencies: services, IAM roles, networking, data flows, SAP-connector configurations
- Designed the split topology – three isolated environments in AWS, separate Kubernetes clusters, isolated Postgres and Redshift instances, separate Terraform state per entity
- Replicated the custom data layer across three environments, including Redshift-specific configurations, custom processing logic, and SAP integration endpoints
- Built validation pipelines comparing data flows between original and split environments, catching value drift before it reached production reports
- Coordinated the cutover: three separate go-live events, each with rollback paths and validation gates
- After the split: stayed on as ongoing maintenance and development lead for the data platform – new processing features, performance optimization
What changed
The split was delivered on regulatory deadline, across all three entities, with zero data loss and zero production downtime exceeding the planned cutover windows.
Ongoing engagement: the engineer continues as technical lead on the platform, currently in month 12 of the relationship.
By the numbers
Hard regulatory deadline with legal consequences; other vendors didn’t want to take the project; our engineer owned the work from architecture to cutover. Reinforce zero data loss / report values correct post-split.
Infrastructure restructuring? You need an owner, not a ticket executor.
Mergers, splits, regional expansion, multi-entity setups - if you need a senior who owns the project end-to-end, this is Dedicated Team territory. Compatible stacks: AWS (Terraform, Kubernetes), Azure, GCP, custom data layers, SAP, enterprise Redshift / Snowflake.