In the middle of the summer, just surfed the web and found an interesting startup company. “Neuromarketing” was the word which grabbed my attention. Just for fun, I applied for the Data-Scientist job (because why not?). After 3 interviews (and tests) I worked my first month with them. 🙂
Before this job, I worked with Machine-Learning and Computer Vision for a little bit more than 1,5 year for a startup company. But there I only learned from the internet because I was the only one who knew these things. It had the good and the bad side too.
The interviews begin…
On one of the interviews where I spoke with the R&D Lead, I got to know how much I don’t know about these things and I just scratched the surface of data-science. But I was really passionate, and I wanted to learn more and more… After the interview I thought that I won’t get the job because I failed on some questions. I didn’t know the answers because I only knew how to use ML and DL for various problems, I just didn’t know the whole background of it, just a little piece.
Now, I’ll skip to the interesting part, enough about the background story. 😉
When I walked to the place at my first day, I imagined to create all the neural nets I dreamed about and I will train the sh*t out of the nets. This was a sweet dream but I started to realize this is not what I signed up for. 😀
One thing I knew for sure: I want to learn everything and even more than that!
On the first&second week I talked with almost everyone in the company and I started to get to know our data. A lot of data (but as we know data is power)… I never saw this many rows and columns in my previous experience. :O
After that first shock I learned about data pipelines and I got a small project: merging data files to create usable data and after that cleaning the data. I didn’t know that it will be so hard because when I thought the data is cleaned the senior members show me 10 other cleaning possibilities. I made some rookie mistakes, like not checking the data type because in pandas DataFrame it looked the same while in one table it was a string and in the other table it was an int.
At first I worked with jupyter-notebooks and after the merging worked there, I had to create pipeline tasks (Luigi framework) so we could save the intermediate files which were needed. After this “session” I started to realize what itvreally means to be a data scientist.
I have to say that it is really interesting because you need a lot of knowledge to create usable data from raw data. And the other important thing I learned in my first 2 weeks: cleaning the data is a lot harder than it sounds and it is a lot of work.
After this time I continued to write the merging tasks and I started to figure out correlation between data (Those who are under 20 are more likely to … etc…). Fortunately I am the full member of the R&D team and they listen to my ideas (so I am not that guy who does the boring monotone work).
So far I really enjoy to work with these problems and I think in the future I will just like it more and more as I will know more and more 🙂
If you have ny questions about the interviews/experience/etc. just write a comment! 😉