How do you ensure that dozens of providers worldwide deliver the same quality? For Mercedes, the answer is simple: PIA! The PIA app, developed by the .msg group through the Romanian dev teams, checks everything from complexity to safety and control in real-time.
The Parts Inspection & Approval (PIA) project has been part of .msg Romania since 2016, when they started developing the CORE part of the application. It's a tailor-made app created for Mercedes providers' network that was ready to go live in just four years. It integrates a very complex process with diverse functionalities, where every error that could compromise the quality of an automobile batch can be detected and traced. In 2024, PIA looks much more different than its first version. Besides the technologies that are constantly upgraded to meet the client's needs and market trends, PIA's ecosystem has grown a lot. Currently, the app is used by 25.000 users from 61 countries, all using the same flow.
How does PIA work?
The application plays a crucial role in the quality assurance, safety, and viability process of all Mercedes automobiles. With the help of PIA, the inspection, testing, and approval processes for each auto part became more efficient and have continuously improved. All plants worldwide that deliver auto parts to Mercedes enter data about every batch, the tests they are running, and their results. This way, conformity assessment & manufacturing data can be displayed and tracked at any moment. Regarding security, the data is secured on different levels depending on the end user.
The team behind PIA
The application is managed by 29 colleagues in Cluj, Targu Mures, and Brasov, three of whom are part of .msg Germany. The team is divided into smaller teams, each owning a submodule of the application. To deliver such a complex application, many technologies are required. The team uses over 15 technologies, from PostgreSQL and Hibernate for databases to Java, Quarkus, RabbitMQ, and IBM MQ for the back end. Since AI is here to stay, Github Copilot is the team's first artificial intelligence tool used in this project. The tools used to manage the Cloud Infrastructure are Kubernetes, Docker, and Helm. There is more to add, but for the moment, we will stop here.
How does the team collaborate?
The Dev teams have to communicate all the time closely, as the application's submodules are interconnected. Everyone on board must be updated with every possible change or upgrade. Before each release, the teams ensure all the necessary tests are run. Any small error can impact the whole process. Right now, the sprints last three weeks. At the end of each one, all releases go live. In the future, the team's goal is to shorten the time needed for the new functionalities to be up & running and to go live with one release per day.
The team's decision to work from the office daily helps the application evolve. Any issue or challenge can be discussed on the spot without needing to schedule a meeting. A big plus is that solutions come from everyone, not only the colleagues involved in the problem.
Another advantage is the collaboration with the client. Annually, there is at least one face-to-face meeting with Mercedes in Romania or Germany, while the conversation is always open online. This is a very productive approach for passionate tech people and the client, who is always keen to see what the team has in mind for their product.