Responsible for infrastructure, designing the architecture and leading a team of developers with the implementation effort of a scalable data-collection system running on Kubernetes and AWS.

Tasks:* Designing a scalable, highly-available, resilient and extensible event streaming platform on AWS* Implementing select components in this platform* Event stream integration with relational data-modelsChallenges included:* Volatile requirements environment* Supporting integration of data lake into rest of the organisation and in the components of the IT landscape

Improved and implemented processes for the integration between IT and business, including, among other activities, processes for code-reviews, requirements acceptance, sprint reviews, business acceptance testing, and stronger stakeholder responsibility into a large (20+) team of developers.

Goal was to replace a real-time legacy API, which existed without any specification or source-code, with a new scalable and robust system hosted on AWS.Tasks:* Delivery Management* Resource Management* Designing system architectureChallenges:* Fully specifying legacy API semantics, requiring reverse engineering* Designing the system to be fully resilient and robust, since there was no budget for any extra post-deployment development or maintenance* Time constraint, requiring trade-offs in regards to different technical approaches

Focus was on weak points in the architecture and the software stack, specifically potential failure points in case of substantial scale-up of business.

Vodefone was releasing the "GigaTV" new generation (cloud, bluetooth, recording, among other features) set-top box together with a hardware OEM with a software release cadence measured in weeks instead of months.My goal was to create an automated black box testing suite in Python, using the HDMI signal as output and a bluetooth adapter as input for the black-box testing, automatically measuring the correctness of each new software release.

Change driver for transitioning from doing releases and deploying data-model schema changes manually in production and other environments to automated and versioned releases using Jenkins and schema migration software. A challenge was getting buy-in and teaching employees with no technical education (many used to work as nurses) to use deployment pipelines and versioned SQL migrations.

* Designing and implementing data integrations from many different datasources of different clinics into a central data warehouse. Each clinic had authority to choose their own software systems.* Building a post-ingestion data-pipeline, involving extensible aggregation for reporting and KPI:s. * Requirements and feedback processes with stakeholders in the hospital operational organization.

Scalable Distributed Computing for Engineering Applications

Spot Detection in Fluorescence DNA Microscopy via Machine Learning

Spetsutvecklingen & prototyping inom .NETHelår 2009. Sedan Somrar + Vintrar.