Software and technology
Develop the tools, integrations and research workflows that connect technical work to everyday business operations.

Project scope
We agree on a reproducible acceptance example and how exceptions will appear to the user. Delivery includes versioned implementation or study files, operating notes and a handover session matched to the client’s technical team.
Describe a representative user task, current systems, data owners, example inputs and expected outputs. A sanitized sample helps us test the process without exposing customer records.
Services
- Custom ERP development
Follow an order from quotation through purchasing, stock movement and invoicing. We define shared records and approval rules, then test the first working flow with the people who operate it.
- SAP integration support
Define which system owns each master record and how errors return to the operator. Coordinate mapping, test cases and reconciliation with the client’s SAP team before enabling an interface.
- E-commerce development
Map product variants, inventory availability, checkout and fulfillment acknowledgements. Test a representative order including cancellation and return so the storefront agrees with back-office records.
- Multilingual website engineering
Design language-specific URLs and navigation around real customer journeys. Static publishing, accessible layouts, search metadata and redirects are tested together before replacing the previous site.
- Mobile business applications
Model the field worker’s task, including poor connectivity, repeated submissions and device changes. Prototype the shortest usable flow and define how the office reconciles synchronized records.
- API and systems integration
Document field ownership, message format, retry behavior and duplicate detection. An end-to-end test traces one transaction across both systems and checks totals after interruptions.
- Web data collection
Choose permitted public sources and define the fields, update frequency and attribution. Build change detection and validation so a source layout change is visible before downstream reports use bad data.
- Robotic process automation
Identify repetitive tasks with stable inputs and explicit exception paths. The automation keeps execution records and a manual recovery step for missing documents or ambiguous approvals.
- Applied artificial intelligence
Select a bounded task such as organizing technical documents. Define a reference set, acceptable error and human review points before comparing model output with the existing workflow.
- Machine learning evaluation
Check how labels were produced and reserve representative data for testing. Report errors by operating condition and compare the proposed model with a simple baseline.
- Full-stack web applications
Connect React or Next.js interfaces to Node.js or NestJS services through explicit contracts. Implement a complete user task first, including validation, permission checks and useful failure messages.
- Legacy software modernization
Inventory C, C++, C#, .NET, VB, VBA, PHP, Java, Ruby or Delphi components and their dependencies. Capture current outputs before replacing modules so important behavior survives migration.
- Server and cloud configuration
Map workloads, availability needs and recovery objectives before choosing hosting. Configure a repeatable deployment and verify that backups can restore a representative workload.
- Database architecture and migration
Trace relationships and business keys in SQL, MySQL or PostgreSQL systems. Rehearse migration with row counts, reconciled totals and an explicit cutover and rollback sequence.
- Research image analysis
Design reproducible Python or MATLAB workflows for research datasets, including CT and MR volumes. Document preprocessing, spatial calibration and validation against reference annotations for nonclinical evaluation.
- 3D reconstruction and visualization
Convert aligned image sequences or measured geometry into inspectable surfaces and animations. Record assumptions about scale, spacing and interpolation so reviewers can distinguish measurements from reconstructed detail.
