Large-Scale Integration Project – Scrum Master / Architect / Fullstack Developer

The project was about restarting and delivering a complex manufacturing system integration after an initial halt. As Scrum Master, architect, and developer, I led the team in connecting ERP/MES, production lines, robotics, and quality systems using an agile, test-driven approach. The result was an on-time delivery with most defects caught early, a flexible message-broker architecture that reduced change costs, and DevOps practices that improved transparency and speed — earning strong customer praise.

S – Situation

The original fixed‑price integration project connecting five manufacturing systems with the customer’s ERP/MES, production lines, robotic automation and quality systems was halted after the requirements phase. It was later re‑scoped and sold under a variable‑price agile model. Functionalities were defined via ATDD tests, and implementation only began after passing acceptance criteria. A five‑day on‑site workshop in North America finalized a key interface specification.

T – Task

I served as Scrum Master, architect and developer and acted as an agile contract model consultant. I led the sprint process with a team of three developers plus myself and worked closely with the customer’s Product Owner to ensure success on both the technical and process fronts.

A – Actions

ATDD phase definitions

  • Collaborated with the customer and MES vendor to write and ratify acceptance tests (ATDD/BDD).
  • Held approximately three remote meetings per sprint (1–2 hours) which brought to light and fixed about 90 % of issues during the ATDD phase; the remaining 10 % were discovered during sprint reviews and testing sessions.
  • Iteratively clarified and updated interface specifications with the MES provider.

Architecture design & Message Broker

  • Designed the integration architecture so that message routing and transformation were handled by a centralized message broker.
  • The broker layer allowed easy addition of special‑case handling during testing without requiring modifications to the customer’s delivered automation systems.

Sprint process & technical implementation

  • Organized the sprint backlog, planning sessions and daily Scrum meetings; the number of sprints and timeline remained on track despite minor spill‑overs.
  • Introduced a mock server to simulate REST APIs, enabling frequent demo and test sessions using realistic production data calls.

DevOps improvements & reporting

  • Proposed and prioritized DevOps practices in the backlog, releasing multiple versions to test and customer environments and laying the groundwork for migrating from Jenkins to Azure DevOps.
  • Facilitated regular test sessions with the MES provider and reported progress, findings and budget status to the customer’s Product Owner and other stakeholders, ensuring transparency.

R – Results

  • Quality & bugs: Around 90 % of defects were found and fixed during the ATDD phase; only 10 % remained to be addressed during sprints.
  • Schedule: Sprints were completed on the planned timeline without significant delays.
  • Flexibility: The message‑broker architecture enabled special‑case handling during testing without changes to the automation systems, reducing coordination and change costs.
  • DevOps impact: Frequent releases accelerated the feedback cycle and established Azure DevOps practices later adopted by the MMS platform team.
  • Customer satisfaction: The Product Owner and other stakeholders praised the clarity and transparency of progress and expressed their satisfaction with the collaboration.

Tools & Technologies

  • Languages & Frameworks: C#, .NET Framework, Angular, .NET 8, Kestrel
  • Messaging: MQTT, Mosquitto
  • APIs & Testing: REST, OpenAPI, Gherkin (ATDD/BDD), OAuth 2.0, Jira, Confluence
  • Mocking & Deployment: MockServer, Docker
  • DevOps: Jenkins, Azure DevOps
  • Database: SQL Server