Data platform engineering and 24/7 operations
We run the platform your business runs on. And turn its data into decisions.
Cloudtech designs, builds and operates enterprise data platforms and AI integrations, securely, from the systems that produce data to the decisions it drives. On premises and under your control, or in the cloud. Around the clock.
- 1996 Engineering heritage Unix engineering since 1996.
- ~30yrs Data management Extensive data management experience.
- ≤1h P1 reaction, 24/7 For every client, every hour of the year.
- 5 Operating regions Europe, the USA, South America, the Middle East and Asia.
01 Service lines
Six service lines. One engineering organisation.
Each service line can be bought on its own. Together they cover the data layer from the first record to the last decision.
-
01
AI integration
Open-weight models on your hardware or in the cloud, or closed models by API, applied where they pay.
- Embeddings on your own data
- Ontology-based full-text augmentation
- Agentic workflows with guardrails
- Decision model training
-
02
Managed platform and database services
The systems the business runs on, operated 24/7.
- Oracle, PostgreSQL, MySQL, SQL Server, time-series
- Failover clusters, Kubernetes, storage
- High availability, disaster recovery, backup
- Upgrades, migrations, licence optimisation
-
03
Data platform engineering
Any source, any target, one governed pipeline.
- Change data capture from databases
- Ingestion from brokers, logs, devices and files
- Streaming and batch processing
- Data contracts, quality and lineage
-
04
Analytics, reporting and decision support
Figures people act on, with the records behind them.
- Data warehouses and marts
- OLAP models and analytics
- Reporting and KPI systems
- Live dashboards and drill-down
-
05
Software development
Java, Python and TypeScript engineering for systems that run on data.
- Stream-processing jobs and integrations
- Enterprise back ends and APIs
- Applications
- Modernisation of legacy systems
-
06
DevOps, automation and observability
Every change versioned, every component watched.
- CI/CD and GitOps
- Infrastructure as code
- Database DevOps
- Monitoring, alerting and incidents
02 Architecture
Any source, one pipeline, decisions on premises.
Data is captured where it originates, moved once through a governed backbone and shaped into the figures people use. Every step runs on hardware you own, inside your network. Your data stays in your hands, behind your own perimeter and access rules, not on servers somewhere on the internet, in buildings you cannot inspect, run by people you did not choose. Where data may leave the premises and the system is not mission-critical, the same steps run in the cloud.
- Capture
Change data capture reads databases through their transaction logs. Connectors and gateways take in brokers, devices, logs and files.
- Stream and process
One backbone carries every record with a registered schema. Jobs validate, join and aggregate in event time.
- Store and serve
Warehouses and time-series stores that feed reports, dashboards and applications.
- Apply AI
Open-weight models on local GPUs read curated aggregates and permitted documents, and answer with evidence or not at all.
- Govern and observe
Schemas, access control and monitoring against a declared inventory cover every layer.
03 Software development
The software that makes a platform useful is written for it.
Products cover the general case. A plant's shift calendar, a connector no vendor supports and an operator's screen are particulars, and Cloudtech writes, delivers and runs that software.
- Java
Enterprise back-end software, APIs and integration services that business systems depend on.
- Back ends and APIs for operational applications
- Integration services between business systems
- Flink stream-processing jobs
- Kafka clients
- Python
AI tooling and data pipeline tooling, back ends, automation and monitoring.
- AI tooling and integration
- Data pipeline tooling with Polars, PySpark and PyArrow
- Django, FastAPI and Flask back ends
- Automation and monitoring
- TypeScript
Used primarily for the front ends of Java and Python back ends.
- Angular applications for dashboards and operational screens
- Live updates over server-sent events
- Role-based access, several languages
- NestJS APIs where TypeScript fits best
- SQL and PL/SQL
Database logic, and the speed of the queries the business runs on.
- SQL and PL/SQL database logic
- Query analysis and tuning, execution plans
- Indexes, statistics and partitioning
- Versioned schema migrations with Liquibase and Flyway
04 AI where it pays
AI is applied where it pays, and nowhere else.
AI is not a goal in itself. There is a lot of hype around it, and most things can still be done without it at a fraction of the cost. We apply it only where it brings a measurable, significant gain in a specific part of a business process.
- Measurable gain first
A model enters a process only when it beats the same process without it.
- Your data stays on premises
Open-weight models on local GPUs. No external service receives a copy.
- The model sees what the role may see
Aggregates, not records, and documents filtered by permission. The input is shown to the user.
- Evidence or no answer
Without evidence in the data, the request is refused, not guessed.
05 Why on premises
Critical systems belong on premises.
We call it the on-prem private cloud: capacity you own, run with the automation and service levels of a cloud.
- Control
Production data stays under its owner's physical and legal control.
- Predictable cost
Capacity bought once and used for years. No metered egress.
- Performance next to the data
Pipelines, stores and models on the same network as their sources.
- Open components
Kafka, Flink, PostgreSQL and Kubernetes. No dependency on one cloud.
RegulationGDPR, NIS2, DORA and the AI Act leave accountability with the organisation that owns the data. On premises does not make anyone compliant. It keeps the answers to where data is processed and who can reach it inside a perimeter you control.
Cloud targetsSelected streams go to Microsoft Fabric, AWS including Amazon S3, Google Cloud or Snowflake when the group needs them. The primary copy stays on premises.
06 Industries
Where the data is the operation.
Where a stopped system is a stopped plant, grid, vessel, hospital or service.
- Manufacturing
Plant databases and machine data streamed into live quality KPIs; failover clusters at plants worldwide.
- Energy
Production and delivery data from electricity, oil and gas, streamed into live operational figures inside the operator's own perimeter.
- Finance
Formal change management, rehearsed recovery, reporting that never queries the core.
- Logistics and transport
Warehouse, forwarding and carrier systems operated 24/7, with the integrations and data warehouses built on top of them.
- Healthcare
Health data platforms where access is controlled and every use is accountable.
- IoT and technology
Operation of open-source IoT platforms for device and sensor data.
- Biotechnology
Genes and proteins designed, read and tested with AI models on the institution's own GPUs; the models grounded in the lab's graph, databases and documents.
07 Service levels
Operated to a service level. Delivered to a scope.
One governance model for a platform run for years and a project delivered in a quarter.
Service levels
- 24/7 operations
On-call engineers every hour of the year.
- P1 reaction within one hour
For every client. An engineer working on it, not an automatic reply.
- Change management
Every production change planned, approved, recorded and reversible.
- Reporting and auditability
Service reports each period. Every change traceable.
Engagement models
- Managed services
Responsibility for an agreed scope, under a service level.
- Project delivery
Fixed scope and schedule, ending in hand-over.
- Professional services
Expert work on demand.
08 Platforms and competencies
The platforms we operate and the tools we build with.
Databases
- Oracle
- PostgreSQL
- MySQL / MariaDB
- Microsoft SQL Server
- QuestDB
Platforms
- Solaris Cluster
- AIX
- Linux
- Kubernetes
- Hitachi storage
Availability
- RAC
- Data Guard
- SharePlex
- Veritas NetBackup
Streaming
- Kafka
- Flink
- Debezium
- Apicurio
- RabbitMQ
Engineering
- Java
- Python
- TypeScript
- NestJS
- Angular
- SQL
- PL/SQL
Automation
- GitLab CI
- Bitbucket CI
- Ansible
- Terraform
- Helm
- Liquibase
- Flyway
Observability
- Zabbix
- Grafana
- Prometheus
- ELK / EFK
AI
- Open-weight models
- Ollama
- Local GPU serving
Cloud targets
- Microsoft Fabric
- Azure Event Hubs
- AWS
- Amazon S3
- Google Cloud
- Snowflake
Certified competence: Solaris on every release from 2.6 to 11, Hitachi Data Systems storage engineering, and Oracle Certified Professional database administration.
Talk to an engineer.
Describe the system, the problem and the time frame. An engineer replies, with questions before proposals.