Martin Svensson, Engineering Director

Leads the product and engineering teams that develop our core datasets and key metrics and safeguarding these from any unwanted behaviour. Our tribe is roughly 50 people including engineers, data scientists, product leads and engineering managers.

Managing a team of engineers that develops and operates Spotify's data infrastructure. Focused on our hiring strategy and growing the organisation to roughly twice its size. During my tenure I also worked as a manager in the Analytics team to help bootstrap their new organisation. Responsible for the technical overview of our platform

Lead research and development in next generation big data technologies, including large scale data bases, analytics and information visualization.

Responsible for starting research activities in Silicon Valley. The work included formulating the research agenda and starting up a new research group that was focused on mobile social media and the interplay between the Internet and Telecom application environments

Ericsson is a world leading software and telecom vendor. Ericsson has over 100.000 employees with a global research and development unit. In my role as a senior research I was responsible for leading our data mining and recommender system research project.

In my role as senior research I lead the social computing group at SICS. The group's main focus was on social and mobile services research as well as recommender systems. I managed projects and did practical work including programming, writing papers and field studies.

Swedish Institute of Computer Science (SICS) is the leading computer science research institute in Sweden. It consists of roughly 100 researchers working more or less independently. As a researcher I worked on information navigation specifically looking at how it could be boosted by recommender system type functionality. This role included developing code, writing research papers, defining research topics, giving presentations and managing smaller projects.

Search and Find developed a high performance search engine for Intranets. At Search and Find I worked a developer with a focus on API development and document filters. Specifically created the Java JNI wrapper towards the underlying C++ search engine.

Mellowtech Core is a developer library that myself and Rickard Cöster started to work on in 2002. Mellowtech Core is a set of components that we use for doing disc and byte based manipulation of Objects. It is typically useful for any scenario that involves storing and sorting large amounts of Objects (in the millions) on disc. Mellowtech allows for consistent and fast disc based object retrieval.

Restlet is one of leading java frameworks for restful web services. It consists of a core part and a number of extension. Myself and a Collegue was the initial developers of the OAuth2 and OpenId extension for restlet. OAuth2 and OpenId are the dominate web mechanisms for doing authentication and authorization.

MellowDB is a hybrid database engine (mixed column based and row based layout) with powerful search functionality based on the Lucene search engine. Performance benchmarks show that MellowDB outperforms the H2 Database Engine which is commonly regarded as one of the fastest embedded databases for Java.

Diary writing is something that many people engage in and is a very important tool to keep track of the important things. In this project we were looking at ways of enhancing the diary writing process by adding contextual and temporal information into the diary itself. The Diary itself was developed using C# on a tablet and Windows Mobile. The project was later commercialized into a product called Affective Health.

Mobitip was an early example (2005) of a location based mobile app for recommending places - such as restaurants and shops. It was developed on Symbian and used bluetooth to automatically exchange recommendations between people that were close to each other. The recommender system itself was a hybrid centralized/decentralized one allowing it to function while not being connected.

Kalas was a very early recommender system - similar to Amazon's: "people who bought this book also bought". The difference being that we created Kalas back in 2000. I developed the back end and client parts of Kalas as part of my PhD thesis work. In this project we also did extensive user studies to find out how and what triggers people to use recommender systems or more generally engage in social navigation. Kalas is the Swedish word for party.