Work Projects

As part of modernaisation work at DFDS we made a number of cloud-native components, however we needed a way to seed the data from the monolith to the new source of truth, and vice versa.

This is where Kafka Connect came in, utilizing a Debezium source to generate messages from the CDC tables embedded within MSSQL. Using a series of open source and a couple of in-house transforms written in Java we could create well-formatted Kafka messages, which then drove a series of components written in .NET and deployed to AWS.

The Kafka Connect infrastructure was deployed using Kubernetes and Flux, and processes hundreds of thousands of messages per location, all day every day.

Before we were given the project, Automotive operations at DFDS relied on a 60 x 10 character terminal-based app last touched in the early 2000s. Because of software incompatibility, devices capable of running this application were decreasing rapidly, and a new solution had to be made in a very short timeframe.

Using the kafka connect setup mentioned here and a series of receivers, APIs, websocket servers and a frontend, a cloud-native application was created from the ground up. This greenfield project allowed for a number of team firsts, including our first react site and our first websocket integration for live updates.

Feedback from our first MVP go-live and beyond has been overwhelmingly positive, with users enjoying the speed and reliability afforded by our modern approach to the new solution, which now processes thousands of cars a week across Europe.

Built upon our legacy software stack, to help provide support and new features for those on the ground. Worked on maintaining and expanding on our product in an industry where every second counts.

• Expanded on functionality involving Mafis
• Added audit history components
• Implemented the display of GPS tracking data
• Allowed shunting from within the app,
• and more…

Flutter frontend talking to .NET APIs and a VB6 backend, with many new features making use of our new components for increased speed and reliability.

I have supported this product on international go-lives both close to home and out in Turkey, as well as fixing production issues out of hours. Support ranged from small tweaks to complex issues with slightly malformed EF queries overwhelming the database server.

Photo taken from Turkey go live, Summer 2024

What did I do?

I was part of the backend development team, working on core aspects such as payment via Stripe Connect and real time chat. The application is still in active development.

What did I do?

I was part of the frontend development team working on key areas such as in app purchasing, calculation of calories for custom areas, as well as UX areas such as filtering and the onboarding process. The backend was a nosql database in firebase, and IAPs were handled via Qonversion. The app is available on the Apple App Store

LINK

What did I do?

I was part of the backend gate control team, working on a variety of areas such as client control apps and software deployment technologies. This involved learning how the tech stack worked, making Azure pipleines and creating lambdas to fit into frontend applications.