I applied online. The process took 3 weeks. I interviewed at Searchmetrics (Berlin) in Nov 2015
Interview
I applied online for a data scientist in the new customer intelligence team. I was selected for a telephone interview, which was conducted by a professional who asked general machine learning questions. I was then invited for a face to face interview in Berlin. They warned me that they won't pay for my trip. I was interviewed by the team leader and someone else. The interview was technical and we had an interesting discussion on machine learning and also some details on the job and the plan set by the team leader who was a professional young guy. I was then selected for the last stage interview. It was with the team leader and one senior manager. He asked me the regular competitive questions. At the ended of the interview the senior manager admitted he had a very positive impression, and I was asked about the starting date and the salary. I was finally told they will get me an offer, and it will take just a week.
The feedback was very positive, and the team leader confirmed that I was the best candidate with strong machine learning skills.
I waited for the offer, but over a sudden I received the traditional refusal email saying unfortunately we had to go for other candidates who's profile matched the job better!!!!
This gave me what company I was going to work for! I am very happy that I did not join them, but sad they I will not be working with that young professional (the team leader)
Interview questions [1]
Question 1
Describe how SVM work?
How can we get probable computed through SVM in a classification problem?
How GP work and how to select its model?
I applied online. The process took 3 weeks. I interviewed at Searchmetrics (Berlin) in Jan 2016
Interview
I was directly invited to a personal interview since I was already living in Berlin at the time. This first interview took place in one of their conference rooms and I was questioned by three members of the data science team. The interview took a couple of hours which was mainly due to lively discussions about the topics. Since I had to leave at some point, a second interview was scheduled where I talked to the same people and lastly had a short interview round with the VP of engineering.
The questions were always fair and I was not expected to know every algorithm by heart, but to understand the core principles. If I did not know the answer to the questions they helped a little and let me derive the answer myself (which did not feel as much like failing). When I had different opinions on topics that was respected as well.
I had the impression that they really wanted to be sure about the candidate. Spending almost 5 hours in interviews was a bit exhausting, especially when being interviewed by multiple people at once.
Two weeks later I got a phone call from an HR employee that made me an offer. We had already discussed salary in the interview but the caller did not know about this. Therefore I had to renegotiate on the phone which was a little annoying.
Still, overall the interview was very fair and the recruitment process painless.
Interview questions [1]
Question 1
- General questions about machine learning algorithms and models
- This is an important one: How do you evaluate a model after training? Accuracy, precision, recall and F-measure
- What can go wrong when training a classifier? Skewed samples, overfitting etc.
- Can you use a random forest classifier to perform regression?
- Explain how SVMs work
- Software development tools (Git etc.)
- Explain clustering by elaborating any one clustering algorithms (e.g. K-means)
