Four areas, one discipline: making data dependable.

Data standards, interoperability and regulatory assessment

When data has to move between institutions, databases and jurisdictions, the format decides whether it works. We design exchange formats, test them against the domain requirements, and build the tools that demonstrate compliance with them.

• Design and specification of exchange formats. XML schemata, ontologies and data models — agreed with expert bodies and documented so that they remain auditable years later.
• Validation tooling. Console, desktop and web service validators so that data providers can check their files themselves before submitting.
• Scientific assessment. Pharmacokinetics and pharmacodynamics from non-compartmental analysis through to bespoke models; QSAR and QSPR models for physicochemical and ADME/Tox properties, including applicability domain.
• Exposure and risk assessment for products and materials, including derivation of permitted limits and preparation for communication with authorities.
• Data curation from publications, trial registries and agency data, with documented provenance for each data point.

Software and platform engineering

Where no product on the market fits the task, we build it. Our projects are rarely greenfield: they are usually systems with a history — existing codebases, data models that have grown over years, several teams involved.

• Domain applications and knowledge platforms for research, public administration and industry — including authentication, permission and role models, multi-tenancy and operational monitoring.
• Mobile applications with location search, map integration and connection to existing branch or inventory data.
• Test automation and quality assurance. We have built test coverage for systems with several thousand methods, including automated consistency checks and crash reporting from live operation.
• Taking over troubled projects. Codebases left by another team, undocumented interfaces, data models without a schema. We start with an honest assessment, not a promise to rewrite.
• Multi-tenant SaaS architecture. Tenant isolation at the database level (row level security) combined with entitlement logic in the application layer — the database answers whose data it is, the application answers what content someone may see.
• Review of other people's data models. We load a proposed schema with the real content at real volume before anything is built on it. Mismatched enumerations, special-case paths and permission rules then surface before the first customer, not after.
• Operations and infrastructure for the systems we deliver, where wanted — server configuration, monitoring, domain and certificate management.

Data science, AI and forecasting

We build models that go into operation and remain verifiable there. That means documented validation, an honest statement about uncertainty, and a handover that puts your team in a position to maintain the model themselves.

• Language and document processing to extract structured information from contracts, reports and technical literature — including building and annotating the training corpora.
• Knowledge assistants over your own content (retrieval-augmented generation): your manuals, standards or technical literature are segmented, embedded and made searchable so that answers trace back to your sources instead of being invented. We work vendor-independently and have migrated such systems between cloud platforms without rewriting the application.
• Machine translation of technical texts with subsequent expert review — for handbooks and training material where terminology and structure have to survive the translation.
• Sales and demand forecasting on sales, production and market data, including scenario analysis.
• Machine learning for classification, regression and anomaly detection, with cross-validation and documented methodology.
• Data preparation and architecture as the foundation: no dependable data basis, no dependable model.
• Visualisation and reporting built for decision papers, not for screensavers.

Interim management and technical leadership

Some undertakings need less additional development capacity than someone who owns the technical decisions and defends them to the board and to the client.

• Technical project and programme leadership for in-house development, licensing and external suppliers.
• Building a product from an idea into something sellable. We accompany young product ventures over several years — from the first assessment through architecture, subscription and entitlement models to launch readiness, including the unwelcome judgement about which features are not yet solid enough to ship.
• Coordination of distributed teams across locations and time zones, with auditable time and effort records.
• Support in expert bodies and liaison with advisory groups and regulators.
• Bid and tender support for public authorities and international organisations.