# Data & Cloud

> Data and cloud engineering from Boltcliq: integration, lakehouses, governance and cloud modernisation with attention to cost, resilience and data residency across Africa.

Canonical URL: https://www.boltcliq.com/solutions/data-and-cloud
Part of: https://www.boltcliq.com/llms.txt

The foundations every intelligent system stands on: governed data on infrastructure that suits the region.
- Data scattered across systems, vendors and spreadsheets cannot support AI. We integrate it, govern it and run it where it should live.
- Integrated: Legacy, partner and new sources in one governed layer
- Cost aware: Cloud spend designed and watched, not discovered
- Resident: Data placed where your rules and customers require
- Portable: Architectures that can move between providers
### What we do
- Every AI project eventually becomes a data project. The model is ready, but the customer table is duplicated in three systems, the branch data arrives as monthly spreadsheets and nobody can say which number is right. We start there. Our engineers inventory the sources, trace how data flows, fix the critical fields and build the pipelines that keep them fixed.
- We then put the data on infrastructure that suits your risk, budget and location. That may be a public cloud region, a private cloud, an in-country data centre run by a local partner, or a mix. We design for resilience in places where power and links fail, for cost in places where foreign currency is scarce and for data residency where the law or the customer requires it.
- Governance is part of the build, not a document at the end. We put lineage, quality checks, access control and retention rules into the pipelines themselves. When a regulator, a board or a partner asks where a number came from, the answer is a query, not a project.
### From scattered systems to a governed layer
- A typical estate includes a core system that is ten years old, several departmental tools, vendor portals and a lot of files on shared drives. We do not ask you to replace any of it first. We connect to it where it is, take what is needed and give each source an owner.
- The new layer is built in stages. The first stage serves one real use, such as risk reporting or a supply forecast. Later stages add sources and users. Every stage leaves the estate easier to run than before.
- Connect first, replace later if ever
- Each source has a named owner
- Each stage serves a real user
### Capabilities
**Data integration**
- We connect core systems, ERPs, mobile money platforms, partner feeds, files and scanned documents into one layer. We handle the awkward cases: batch exports, flat files, APIs without documentation and systems that cannot be queried at peak hours. Each pipeline is tested, monitored and documented.
- Connectors for databases, APIs and files
- Change data capture where possible
- Documented and monitored pipelines
- Nothing intelligent happens until the data arrives.
**Lakehouse and warehouse**
- We design the storage and processing layers for analytics and machine learning, using open formats so you are not tied to one engine. Raw, cleaned and curated zones are separated, and costs are tied to usage. We size the design to your volumes, not to a vendor reference.
- Open table formats
- Raw, cleaned and curated zones
- Right-sized compute and storage
- The right size saves more than any discount.
**Cloud modernisation**
- We migrate and re-architect workloads for security, resilience and cost. That can mean lifting a system as it is, rebuilding it as services or retiring it. We write a plan with sequencing, rollback and a cost view before moving anything.
- Assessment and sequencing
- Landing zones and security baselines
- Rollback and cutover plans
- A migration is a risk project before it is a technology project.
**Data governance**
- We put in place the catalogue, lineage, quality rules, ownership and access policies. Sensitive fields are classified and protected. Policy is expressed in code so it is applied the same way in every pipeline and every query.
- Catalogue and lineage
- Quality rules and ownership
- Classification and access policy
- Boards and regulators ask where numbers come from.
**Residency and sovereignty**
- We map where each class of data may be stored and processed and design the architecture to match. Options include region selection, in-country partner facilities, encryption with keys held by you and controls on cross-border flows. Which laws apply depends on your sector and countries, so confirm with counsel.
- Data classification by location rules
- Customer-held encryption keys
- Controls on cross-border transfer
- Location decisions are cheap on day one and expensive later.
**FinOps and reliability**
- We tag, budget and monitor cloud spend per team and per workload. We design for failure with backups, tested restores and sensible redundancy for sites with unreliable power. Reports show spend against value, so reductions come from design rather than panic.
- Cost allocation and budgets
- Backups and tested restores
- Resilience for unstable power and links
- Currency swings make surprise bills especially painful.
### Where it applies
- Banks and insurers
- Customer data sits in a core banking system, a card platform, a CRM and several branch spreadsheets, and regulatory reports take weeks.
- We build a governed layer with lineage, so reports are generated from the same cleaned data and every figure can be traced.
- The head of data owns definitions. Compliance decides what is reported and how.
- Telecommunications
- Network, billing and mobile money data are huge in volume and held in separate platforms.
- We design a lakehouse with cost controls, tiered storage and access policy for analytics, fraud detection and network planning.
- Data owners decide who can access subscriber data and for what purpose.
- Government and public sector
- Ministries and agencies hold records in separate systems, and citizen services require repeated paperwork.
- We build secure data exchange between agencies with consent and audit, hosted in-country.
- The agency and its legal advisers decide what may be shared and under which law.
- Healthcare
- Hospitals and clinics hold patient, stock and billing data in different systems with uneven quality.
- We integrate records with identifiers, consent and access controls, and give managers reliable operational reports.
- Clinicians and privacy officers decide on access and use of patient data.
### Built on
**Argus**
- Provides the private data and cloud platform, on Postgres on AWS first, portable to other regions and providers.
- Boltcliq product
- Data stays under your control.
**Zeus**
- Adds the shared model and governed metrics on top of the integrated data.
- Boltcliq product
- Cleaned data gains business meaning.
**Odin**
- Controls who and what can reach each dataset, for people, services and agents.
- Boltcliq product
- Access policy is enforced in one place.
### What we do and do not do
- Integrate and clean data from your existing systems
- Design and build lakehouses and pipelines
- Plan and run cloud migrations in stages
- Embed governance in the pipelines
- Design for residency, cost and resilience
- Insist on replacing working systems
- Resell cloud capacity as our business
- Move data across borders without your decision
- Hide cost behind long contracts
- Offer legal advice on data law
### Questions we hear
- Do we have to move to the cloud?
- No. We design hosting around your constraints. That can be public cloud, private cloud, an in-country facility or a mix. We help you compare cost, risk and rules.
- Which cloud providers do you work with?
- We work with major providers and with in-country operators. Our platform runs on Postgres on AWS first and is built to be portable. We do not recommend a provider before we know your needs.
- How do you handle data residency?
- We classify data by the rules that apply and place each class accordingly, using regions, in-country partners and customer-held keys. Which laws apply depends on your sector and countries, so confirm with counsel.
- How long does integration take?
- It depends on the number and condition of systems. We deliver the first governed output within weeks, then add sources in stages.
- What happens to our legacy systems?
- We connect to them first. Replacement is a separate decision, made when the business case is clear.