Daniel Köhnlein

This platform replaced an ageing SAP system. It ingests data from external systems, transforms it, and serves it back out to everything downstream. We built all of it on AWS, infrastructure included.

Performance drove the design, because millions of records have to move through it promptly. The harder problem was the interfaces at both ends: a mix of technologies, delivery patterns and ages, every one of which had to keep working.

I sat between engineering and delivery. As technical lead I gathered and documented requirements with the project owners, the business and the product owners, then cut and scheduled the work in refinements. Ticket quality stayed high and the pace stayed steady.

I ran the meetings and kept them short and on point, advised on technical decisions, and made sure the right people were in the room and actually got heard.

I also prepared and documented the codebase for its security clearance, and worked on the infrastructure side: automating deployments, tidying up the infrastructure code and extending it. The platform has shipped and is running.

The company had accumulated a lot of overlapping software. This project set out to bring those functions into one place: show the recorded data, analyse it, connect devices to feed it, and configure and control those devices from the same system.

I did the early architecture work and proved it out in a proof of concept. Components talked to each other over a message bus to keep the coupling loose, which is what the flexibility requirements demanded. Holding data centrally while still surfacing the relationships across it led to some promising work with document and graph databases.

It was early enough that it never reached the market, which was expected. The ideas landed well, though, and are still being taken forward.

This started as an external engagement, advising on how to line the product up with where the company was going technically. I reviewed the architecture, assessed it, and worked through the recommendations with the external team. That review became the basis for everything that followed.

Bringing the software in house was always the plan. I worked out the requirements with the product owner and put together a plan for the integration.

I converted the existing pipeline to the company's own tooling, then coordinated building out the infrastructure with the product owner, including a Kubernetes cluster and the container registry to go with it.

Once that was up I wrote the Helm charts for the pipeline and the deployment, and rewrote the test pipeline to run inside the cluster.

The product shipped. By the time I left, customers were using it, development was ongoing, and the move onto the company's own infrastructure was nearly done.

This began as bought-in software that my team and I took over.

I started on maintenance and new features, which gave me the time to learn the stack properly. That turned into a run of architectural improvements, in the software itself and in the development environment, the pipeline and the deployment.

Development leadership moved to me off the back of that. My manager and I worked on the process, introduced Scrum, and did the groundwork for the Atlassian suite the company now uses everywhere. I took the Scrum Master role: keeping the process honest, looking after the team, the product owners and the stakeholders, and handling prioritisation and release dates with the product owner.

I took Git in hand as well. Introducing Git Flow separated development from releases cleanly, so we could keep building while a release was being prepared.

When QA came to me most of the testing was manual. We automated it and reached 98 percent coverage of the testable frontend, which meant we could finally say something meaningful about quality. Confidence in the product went up across the group.

We took unit test coverage to 80 percent, with the same effect: real confidence in what we were shipping. That showed in team morale, in how the product owners and stakeholders talked about the product, and in the end in what customers thought of it.

The product sells today and runs at a range of customers who are happy with it. After I left, my team carried on maintaining and developing it.