“Machine Learning retrieves the unusual out of the ordinary”
An interview with our ML-expert Teun

Blog Taavi


By Qlouder


Teun Krikke is one of our ML experts.  He works in close cooperation with our team of cloud experts on applications using Machine Learning and specifically Deep Learning. An interview with Teun about his job at Qlouder and his vision on Machine Learning.




How did you end up at Qlouder?

I work (remotely) part time for Qlouder. I’m also working on getting my PhD at the University of Edinburgh. During my Software Engineering studies, I started programming fanatically. During an internship in Australia, I learned that I find troubleshooting more interesting than ‘just’ building an application. After returning from Australia, I still had a few years of college left and I wanted to gain more experience. I ended up at e-office and worked with Stefan who, at that time, lead the Google branch at e-office in Houten. After my internship we stayed in touch. In 2012 Stefan founded Qlouder and he asked me straight away to join the team.

The potential of Machine Learning is enormous. What are your thoughts on this?

The capabilities of Machine Learning are almost infinite. You can apply it in everyday applications, like Facebook. Tagging someone’s face for example is one of the many forms of Machine Learning. But also in terms of security Machine Learning provides added value.


In my opinion Machine Learning retrieves the ‘unusual out of the ordinary’. Security cameras at an airport for example.  A camera monitoring a high fence to protect a runway. This camera records video-images 24/7. A Machine Learning application can be taught to categorize situations in (as an example for this situation) three components: plane spotters on allowable distance from the gate; plane spotters on unauthorized distance from the gate and someone with a hedge trimmer at the gate. When an application quickly and easily identifies the person with the hedge trimmer from all the video-images, and can also report it to the security at the airport, it immediately increases safety.


In which other ways offers Machine Learning added value?

Any organization can apply Machine Learning. Screening keywords in a text for example can be performed by an employee up to 10 hours a day. The downside is that an employee gets tired. A Machine Learning application never gets tired and can continue screening texts 24/7.


Do you think that certain jobs may disappear because of Machine Learning?

Yes, some jobs will. However, Machine Learning itself learns nothing. There are always algorithms which can be faster. You need people for that. In theory it is of course possible that applications become so smart that they eventually replace everything. But we are definitely not at that point yet.


What are you currently working on that involves applying Machine Learning?

For Qlouder I work as a consultant for a client for whom we develop an application. An application is divided into small parts. For this application, I focus on a specific Machine Learning component; identifying and finding keywords in text or categorizing texts. My colleagues at Qlouder ensure that the application finds certain texts and relevant articles related to the industry of the customer by use of these keywords. It is interesting to be involved in a project in which we use advanced technology; this makes the work at Qlouder more interesting for me.


Which industry adopts Machine Learning the fastest?

Industries like Defense for security. Also, the gaming industry uses a lot of Machine Learning. But in the administrative sector it is often unclear what could be the benefits of Machine Learning. I foresee opportunities for an application that searches for keywords.


How do you stay up to date in your field?

Within the field of Machine Learning I focus on Deep Learning; this means teaching a machine to recognize the difference between actions and gestures. Deep learning can be applied to video and audio. Because of my job at Qlouder and the extent to which we are already applying Machine Learning I keep abreast of the latest developments. My studies and reading a lot of scientific articles is also a must to keep up in this field.


What is your prediction for Machine Learning in 2020?

Machine Learning becomes part of our lives almost imperceptibly. First mostly on (smaller) devices such as phones, smart watches or the – on Google I/O announced – Google Home. When Machine Learning gains more publicity, organizations will start using it actively. The best thing is the self-driving car, or at least the choice to use it. I expect that the generation that learns how to drive a car now does not easily want to give up control. However, generations after that will consider self-steering cars as obvious.