Black Dog Labs

Case Study

General Register Office

Government

Cloud migration and AI-based registry data classification

Black Dog Labs helped the General Register Office move off on-premise infrastructure onto Azure and AWS, and classify registry data using PyTorch, RAG and OpenCV.

Black Dog Labs ·

Client

General Register Office

Industry

Government

Services

Cloud Architecture, Machine Learning

Tech stack

AzureAWSAWS RDSAWS AuroraAWS KinesisKubernetesMS DynamicsGrafanaJavaPyTorchOpenCV

Black Dog Labs helped the General Register Office (GRO) migrate off on-premise server racks onto Azure and AWS, and introduced AI-based classification for registry data, taking the work from problem statement to a first production version within 12 months in a highly complex regulatory domain.

The challenge

01

Infrastructure tied to on-premise server racks

Core systems ran on physical server racks, limiting the organisation's ability to scale, modernise, or adopt cloud-native tooling.

02

Registry data not classified for modern use

Registry records existed in forms that made them difficult to search, structure or use in downstream services, with no automated way to classify them.

Every change also had to hold up in a highly complex regulatory domain, where registry data carries legal weight and errors are not tolerable.

The solution

The work split into two tracks: migrating the underlying infrastructure to the cloud, and introducing AI-based classification for registry data.

Track 1

Migrating off server racks onto Azure and AWS

Black Dog Labs migrated GRO's infrastructure off physical server racks onto a combination of Azure and AWS, integrating with MS Dynamics and standing up managed data services.

  • AWS RDS and AWS Aurora for managed data storage
  • AWS Kinesis for data streaming
  • Kubernetes for deployment, Grafana for observability
  • Java-based services integrating with MS Dynamics

Track 2

AI-based classification of registry data

Registry data was classified using PyTorch models, retrieval-augmented generation over dedicated RAG databases, and OpenCV for image-based records.

  • PyTorch for classification models
  • RAG databases for retrieval
  • OpenCV for image-based registry records

The outcome

The programme went from problem statement to a first production version within 12 months, in a domain where regulatory complexity typically slows delivery. This describes time to a first production version rather than a measured business outcome; no independently verified savings, accuracy or ROI figures are claimed here.

About Black Dog Labs

Black Dog Labs designs and builds cloud and AI solutions for organisations operating in complex, regulated environments. For the General Register Office, we combined a cloud migration with AI-based data classification, delivered against a demanding timeline.

Discuss your infrastructure modernisation with Black Dog Labs

← Back to all case studies