Interviews I Discontinued or Got Rejected as Data Scientist or Analyst
Now when I am scrolling sometime on the LinkedIn it seems like it have gotten some other social media type virus. I can see people are thrilled, very excited or super excited to join the X company as Y specialist etc. I guess also LinkedIn encourage some of those activity such as KUDOS culture. You might think I have negative perspective on this but what I think personally is, that will lead some weak mind individual to a corporate Narcissist mind set. That convinced me to write that story, which I never thought I would write one day. So I am going to write about my last job search period which happened from May/February 2020 to February 2021. All the rejection and discontinuation I had and in short how they happened and all my mistakes in those interviews , also from which I learned the most and the interviews where there is learning lesson for the candidate after getting a rejection.
Before I start I would like to talk a bit about that kudos culture and extra excitement. I have seen my friends posting stuff in social that how happy they are and suddenly after N month I do not see it any more. Then suddenly I got the new dude we broke up. This is exactly can happen in corporate, as the average stay in a job is 3–5 years but might vary in different regions, the person I will kudos today might be joining my competitor company in 3 years, will you be able to kudos then ? Rather its better to keep the valuation of an individual internally and make C level aware of the importance of that particular individual.
Lets start, these are the companies I rejected from the first screening, like I go some sort of emphatic (you are really good but we can not take you type of email) email. So these have no story to tell I just got rejected on the face. Cleo AI, Mapp Digital, Sixt, IDC, Dunnhumby, MSD, Zeta Global, Lego, IBM and few others I did not remember.
Then I have got few where I passed the first screening but got rejected or after the phone call I found out its not a position suitable for me. Facebook, Payever and few recruiter based agencies are just trying different personal to see which one will fit I guess.
Now here is few interesting story to tell about the companies I went bit far in the interview process. For most of the company I think I did not meet their expectation but also for few of them I think its was wasting of time and should never continue these kind of interviews.
Revolute- They approached to me for a position which was not exactly data scientist or analyst but they call it Data Ops for Product Owners something like that. After first and second screening they sent me a task about fraud analysis and I took data scientist approach to resolve the problem, trying out some models and exploratory analysis, but got rejected.
So from the rejection email I realize actually this position does not need much technical data science or analyst skill but something else probably which I do not have I guess. But I am not interested to have that skill actually, I want to be remain technical for now. This was my first fraud analysis in real so there was something to learn about time series and log data actually.
Sinequanon(SQN)- After second round with the discussion with on of the co-founder I got a nlp task which I had to finish in 8 hours (not strict). I decided to spend exactly 8 hours to train, test, validate a model and deploy the model as web service. That time I just had experience on neural network only for Computer Vision but not RNN or other new technique. So straight I thought I will implement a ml model with scikit-learn and do some grid search. But I spend a lot of time on resolving a problem which I was not aware about:
When saving a scikit-learn Pipeline object using pickle there is some thread issues and I could not resolve that in time. The data set was imbalance and I did not up sample it properly.
So the project was half done, no neural network implementation, no serving model and I got rejected. But they told me they like my python skill and coding design, that was an encouragement. So next day I fixed the pickle issue and was able to learn google cloud run and deploy my project with docker and cloud run. Though I got rejected I was able to improve my knowledge through the interview and I enjoyed the process because the reviewer told exactly what they did not like.
SmartSteel- Position was mix of API developer and data science. So after second round they send me a task to build a REST Api service using flask or django. I knew flask from Udacity Full Stack Nanodegree. So I was very confident. I finished the task and was able to send them properly this time. They replied my work did not fulfill their expectation.
which made me more interested in flask and also made me realize how weak the certification courses are in online platform. Later I learned about moduler rest api development with flask-restful, blueprint, (Un)Marshalling json data, Flask Application factory design pattern. As a result when working on one python Dash framework project at my job I was able to create a new product where I was able to connect multiple dash server under one flask server; a unified place to catalog multiple dash apps which I called DashEngine, you can read more detail in github.
When all these happening I already started learning Go, Julia and Rust and I was becoming more interested in Software Development. But working as data analyst for 5 years I was able to bring those experiences in software development and that was making me a versatile professional who understand analytics, data science, data engineering but also can imagine how software engineering world would looks like.
Concur(SAP)- I had total four round of interview for ML engineer position. All of them are face to face discussion but no given task. I would avoid this kind of interviews when ever possible, I will tell you the reason why. I have been asked several different question on Statistics, Software engineering etc. After 4 round of oral (3 in Prague, 1 in USA) interviews, I suppose to have a call with the department head from USA. But for some reason It was postponed and later I got a rejection email and they say they prefer other candidate.
Now as you can see from that interview I just know that there was someone better than me, but there is zero benefit for me. I did not get any any task to do and it did not help me to learn something new. Rather I could spent these hours on some open source projects or my own learning. So I think these kind of interviews are not much beneficial for candidates where you do not have any scope to show that how much good you are in hands on tasks or if you are not good atleast you will know it.
There is a twist in the story, after 2–3 month I got a call from concur for the similar position again for the same team. But I was not much interested, so I deliberately ask for more salary and they were fine. But they wanted to start again from the first round and I refused, I did not wanted to do the same thing again, what if I got rejected again in the last round?
TechData- This one is same like concur they just asked me to explain my experience with no hands on task. There was first 3 round of interview. The 3rd on is in the office and this time there is a HR person to ask me some question. I thought It was positive, but after I came back home I got email that, they need references from my previous colleagues at-least 1 from each last 2 jobs. I managed to get my last manager from Amazon and one of my colleague from McKinsey . But then they ask me any colleague is not sufficient but they need to talk with the manager. I asked my McKinsey manager but he is in USA. I shared his contact as my manager agreed to talk. For 1 month I did not get any update, later the recruiter informed me that they can not decide between me and other candidates still. So I backed off and ask to inform techdata to pick someone else.
As you can see again the amount of time I spend in that process has no benefit for me, I do not learn any thing new. I did not had an opportunity to show If I am good how much good I am in hands on analytics or even if I am rejected I would atleast know that I have to be much better.
Wrike- I had 2 round of face to face interview, then I had an analytical ability test exam, 20 minute 40 question. I instantly start to remember those painful days for the preparation of my GRE exam 7 years ago and I definitely do not wanted to do that again but the recruiter told me to just go for it. But I did not spend much time on that preparation and I failed in the exam and got rejected. But I do not regret but rather happy that I did not spend many many hours on solving pattern matching, synonyms or high school math. I will tell why?
I already have 4 and half years of experience of working on some of the renowned companies. From my experience I can tell, In my everyday job no one is going to ask me to calculate father or son’s age or how much time it need to fill up a tank of water etc. There is no benefit of learning those for me at that point. They even advertise that their CEO have taken that test, but I do not see any benefit for experienced candidate as there is thousands of companies looking for talent and they are just one of them and there is millions of candidates are looking for job and I am just one of them. They say that they want to check if the candidate can adjust with pressure condition, what does that mean the people are working there are in pressure condition or in stress?
Rather I can spend these 20 hours of preparation time on some open source projects, update some open source documents lets say pandas doc or scikit-learn doc or my own projects and actually by that time I was writing a Julia package Twitter API client written in Julia (TwiliteTimeline) and one project management package written in Julia (ProjectFlow) and I am happy about it that I spend more time on those.
So that interview might lead to a pure wasting of time, with out learning anything and I am happy that I guess it before and did not spent much time on it. There is nothing to learn from these kind of interview.
At that point I have gained a lot of skill personally along side doing analytics with python, pyspark, sql. I was providing sql training, python workshop at my previous job in Socialbakers. I have learned serverless technology, GCP cloud run, Go web dev. docker and I have written 2 packages in Julia and I also have published a Julia package development documentary in my Youtube channel.
Finally I was getting closer to my goal after 9 to 10 month of searching job. Now along side with analytics and data science I know some technology stack which help me to stand out from the crowd. At the end of January after gone through one of the toughest interview task I got accepted as Data Scientist in a company (I will tell you the story another day) and I would definitely miss my previous team at Socialbakers, it was a fantastic place to work. I actually had two fantastic interview, I discontinued one of them as I already excepted the offer from the other one.
So I am not excited or trilled to join some where new but I am focused and steady toward my goal at my new role.
As conclusion thought, the interviews are the best if there is some opportunity to learn even if you got rejected. So I think if you have enough patience to find a job that suits you, you should be concerned about your time and you should also grow that confidence to say no to some interview. Believe me you will not miss any thing rather will save energy and time. When going through those interviews there were few interviews where I back off from the interview and I was able to say no because I figured out my time is valuable too.
