Services
Build. Optimise. Govern. Embed.
Four pillars rather than a catalogue. We would rather do a small number of things to a standard we can defend than list everything and specialise in nothing.
Build
Data platform engineering
Cloud data platforms from scratch or from a legacy stack. Warehouse, ingestion, orchestration and transformation, delivered as reproducible infrastructure-as-code.
Who it is for
Companies whose reporting runs on spreadsheets, an ageing on-premise database, or a cloud setup assembled ad hoc.
Typical duration
6–14 weeks
What it includes
- Warehouse design and setup, sized and structured against your actual workload
- Batch, event-driven and streaming ingestion — managed connectors where they fit, custom services where they do not
- Layered transformations in version-controlled SQL, using a medallion architecture
- Orchestration with dependency management, retries and alerting
- Infrastructure as code, so environments are reproducible and changes reviewable
- CI/CD pipelines for reliable, repeatable deployment
What you get
- A running platform in your own cloud account
- Infrastructure-as-code repository
- Transformation project, versioned and tested
- Orchestration definitions
- Architecture document and runbook
Build
Cloud and platform migration
On-premise and legacy warehouses moved to the cloud — including migrations that have already stalled and need finishing.
Who it is for
Teams carrying a legacy warehouse, or a migration that reached sixty per cent and stopped.
Typical duration
3–9 months
What it includes
- Dependency and lineage mapping of the existing estate
- Target architecture and a cost model you can take to finance
- Phased cut-over with parallel running
- Reconciliation testing, old system against new
- Decommissioning plan for what is left behind
What you get
- Migration plan with phasing and risk register
- Cost projection
- Migrated platform
- Reconciliation evidence pack
Optimise
Cost and performance optimisation
Cut warehouse spend and query latency. Partitioning, clustering, storage tiering and pipeline redesign, with the savings measured before and after.
Who it is for
Anyone whose cloud warehouse bill has become a line item people ask about in board meetings.
Typical duration
2–6 weeks
What it includes
- Spend audit broken down by query, table and team
- Partitioning and clustering redesign
- Storage tiering and retention policy
- Materialisation strategy — what to precompute and what to leave
- Pipeline consolidation and warehouse sizing
What you get
- Audit report with itemised, costed savings
- Implemented changes
- Before-and-after measurement against your actual bill
Govern
Governance, privacy and quality
Lineage, data quality, cataloguing and privacy controls. The register-and-lineage groundwork every audit and privacy regime asks for, built inside regulated banking.
Who it is for
Regulated businesses, companies approaching an audit, and any team that cannot answer "where did this number come from?"
Typical duration
4–10 weeks
What it includes
- Data catalogue and information asset register
- Column-level lineage
- Data quality testing and alerting, wired into CI
- PII classification and handling
- Privacy-regime alignment, including GDPR where you serve EU customers
- Access control and role design
- Retention and deletion paths that actually execute
What you get
- Information asset register
- Lineage documentation
- Quality test suite running in CI
- Policy and procedure pack
Embed
Fractional data engineering
Senior capacity on a monthly retainer, for teams that need the seniority without carrying the headcount.
Who it is for
Scale-ups with analysts but no data engineer, teams between hires, and platforms that need an owner.
Typical duration
3-month minimum · 4–10 days per month
What it includes
- An agreed number of days per month
- Platform ownership, with on-call by arrangement
- Code review and mentoring for analysts moving into engineering
- Roadmap ownership and prioritisation
What you get
- Monthly written report
- Continuous delivery against an agreed backlog
Process
How an engagement runs
- 01
Assess
One to two weeks. We map the current stack, costs and failure points, and hand you a written findings document with a prioritised roadmap. Fixed fee. Yours to keep whether or not we continue.
- 02
Design
Target architecture, migration path, cost model and delivery plan — agreed before a line of production code is written.
- 03
Build
Delivered in increments, in code, in your repositories, from day one. You see progress weekly, not at the end.
- 04
Hand over
Documentation, runbooks and a working session with your team. No lock-in by obscurity.
Engagement models
Start small
Most clients begin with a discovery sprint. It turns an open-ended consulting spend into a small, defined purchase — and it means the project that follows is scoped against evidence rather than assumption.
Discovery sprint
Fixed fee · 1–2 weeks · written roadmap
A first engagement. Low risk in both directions.
Project
Fixed scope · milestone-billed
A defined build or migration with a known end state.
Retainer
Monthly days · 3-month minimum
Ongoing platform ownership and continuous delivery.
Tell us what is breaking.
Thirty minutes, engineer to engineer. If we are not the right fit we will say so on the call and point you somewhere better.
We reply to every enquiry within one working day.