IgaUga.EU
Selected work

Systems we have designed, built and kept running.

A selection of engagements across MedTech, industrial energy, public administration and research. Client organisations are described by sector; references are available on request.

Challenge

Battery systems deployed on vessels generated valuable performance data, but nothing connected them to the company. On the factory floor, cell and pack cyclers and lab equipment produced test data that never left the acquiring device. Field service relied on ad-hoc analysis cycles. Vessels spend long periods offline, and operating cost had to stay low.

What we did

  • Designed a manufacturing data collection architecture from cyclers and lab equipment to cloud databases, with interfaces and scripts for hardware that exposed no native export, and web-based review and reuse of the collected data.
  • Built a secure MQTT-based toolkit that turned deployed batteries into connected devices, with emphasis on security, availability and low operating cost.
  • Designed the on-site data submission component so that offline periods are absorbed and data arrives coherent once connectivity returns.
  • Delivered a remote monitoring platform: near-real-time performance data, instant event notification to service staff, history review, analytics, data conversion and export to external tools.
  • Built an upload-and-analyse web tool for field service, so staff submit collected data and are presented almost instantly with processed, decision-relevant information.
  • Prototyped the replacement of a Windows-based on-site controller with a Raspberry Pi platform offering networked, secure remote control and monitoring, proven in the field before wider rollout.
  • Introduced Agile and DevOps tooling (Jira, YouTrack, Upsource), designed long- and short-term iterations and trained a remote embedded-development team to work against them.

Outcome

Manufacturing, test and in-field data now flow into one platform, service decisions are made on processed data rather than guesswork, and the embedded team runs on a visible, iterative delivery process.

  • MQTT
  • Raspberry Pi
  • Python
  • Cloud databases
  • Remote monitoring
  • Scrum / Kanban

Challenge

A company serving both clinical and commercial users needed a web solution to support diagnostic processes, with methodologies for hip-implant design in human and canine applications integrated into the same workflow, and with the software lifecycle discipline a medical product demands.

What we did

  • Lead the architecture and hands-on implementation of the diagnostic application: structured case data, role-appropriate views, validation and deployment across development, test and production.
  • Integrated implant-design methodologies into one application surface so users see a single workflow instead of disconnected tools.
  • Introduced and operate software lifecycle practices: backlogs, versioning, work tracking, change management, CI/CD and automated testing.
  • Automated software production through GitHub Actions with vulnerability monitoring.
  • Manage cloud deployments on AWS and DigitalOcean, including private VPS provisioning, environment separation and security monitoring.
  • Iterate capture, review and deployment workflows directly with clinical and commercial stakeholders after each demonstration cycle.

Outcome

A cloud-hosted diagnostic platform in production, with research, commercial and engineering work running on the same Agile cadence and a release path that is automated, tested and monitored.

  • Python / Django
  • PostgreSQL
  • GitHub Actions
  • AWS
  • DigitalOcean
  • Security monitoring

Challenge

A software project transferring data from cardiac-device programmers, principally pacemakers, to a cloud data-management platform had been damaged and was blocking the product. It had to be recovered quickly and left in a state that allowed safe, continued iteration.

What we did

  • Recovered and rebuilt the damaged project in record time.
  • Designed and developed the wireless medical-data transfer toolkit between device programmers and the cloud, running on Raspberry Pi and related Linux services.
  • Designed integration and validation steps so clinical data hand-offs remained traceable through the stack.
  • Introduced automated regression testing for the web components, CI/CD and static code analysis to restore a controlled release path.

Outcome

A working, tested data pathway from device to cloud and a delivery rhythm the team could sustain.

  • Raspberry Pi
  • Linux
  • Python
  • Regression testing
  • CI/CD
  • Code analysis

Challenge

Financial reporting depended on spreadsheets exported from the federal SAP system, aggregated by hand. Administrators needed streamlined, graphical reporting and a way to learn about technical problems before end users did.

What we did

  • Led the design and implementation of two web-based financial reporting systems.
  • Built the transfer from SAP sheets into a PostgreSQL database, with aggregation and graphical representation for streamlined reporting.
  • Documented data flows and roles with process owners and iterated the reporting views on user feedback.
  • Launched a real-time error reporting system so technical issues could be corrected transparently to end users.

Outcome

Spreadsheet reporting replaced by production web workflows, including a tool used weekly by the head of financial administration at an embassy.

  • Python / Django
  • PostgreSQL
  • Matplotlib
  • LaTeX
  • Linux / Apache

Challenge

One organisation ran its websites on shared hosting with no central identity or network design. Another kept an archive on a dispersed, mismanaged multi-host system whose security had lapsed. Both needed controlled, documented infrastructure on a non-profit budget.

What we did

  • Led web development, IT support and infrastructure management; maintained web servers and applications.
  • Designed the internal network, built an Active Directory domain, centralised web solutions and managed servers with VMware.
  • Moved the organisation's websites from shared hosting to full VPS automation.
  • Procured large software packages through educational and non-profit programmes, and supported the acquisition of a 50K USD Microsoft grant on the strength of the organisation's software, network centralisation and user management.
  • Renewed system security and installed a new archive server, restructuring the dispersed estate into a concise, controlled data solution and rebuilding data acquisition and storage with minimal resources.

Outcome

Operations became documented and controllable, security was restored, and infrastructure costs were reduced through automation and non-profit licensing.

  • Linux
  • VMware / Hyper-V
  • Active Directory
  • MySQL
  • Firewalls
  • VPS automation

Challenge

Statistical models for detecting disease in medical images, and later for detecting fraud in payment transactions, had to move from research into software and hardware products that clinicians and investigators could rely on.

What we did

  • Designed Markov Random Field and other statistical models for recognising structure on brain-surface images and detecting breast-cancer symptoms in mammograms, with clustering and registration methods for multi-modal historical datasets.
  • Designed statistical models for prostate-cancer detection on ultrasound and managed the computing cluster used for heavy training.
  • Acted as liaison between data research and software development to assure correct, detailed knowledge transfer into a market-ready medical product.
  • Supported statistical models for credit-card fraud detection; coordinated data and knowledge integration across departments; introduced Scrum, test-driven development and continuous integration into research work.

Outcome

A peer-reviewed publication in Computerized Medical Imaging and Graphics (2008), three patents on ultrasound resampling, medical-image presentation and digital medical-image segmentation, and models that reached commercial products.

  • Statistical modelling
  • Markov Random Fields
  • Image processing
  • Matlab / Python
  • Compute clusters
  • Jenkins CI

Working on something similar?

We are happy to discuss how these patterns apply to your system, and to put you in touch with references.

Contact IgaUga.EU Browse services