About Coditation : A product development and solutions firm with decades of collective experience of applying emerging technologies – machine learning & AI, big data, cloud and mobile – to build cutting edge software.

 

Interviewing for Role : Data Scientist

 

Key Skills Required : Python, EDA, Statistical tests, Time Series forecasting, CNNs, RNNs, LSTMs, Data Mining, Inferential Statistics, Regression and Classification algorithms, Unsupervised algorithms, OOPs

 

Round 1 | Telephonic Interview with HR

An HR round where she gave an introduction about the company, the type of person they are looking for and my availability dates. It was a screening round.

 

Round 2 | Telephonic Interview with Data Scientist

It was round focussed on data science projects of my past. The interviewer went into depth of the image classification project and asked questions like:

(1) What were the pre-processing steps?

(2) How was benchmarking done?

(3) Model selection and hyper-parameter tuning

(4) Steps taken to tackle class imbalance

(5) Could GANs be used to address class imbalance?

 

Round 3 | Google Meet Interview with Data Scientist

Started with a brief introduction to Coditation and himself, and then discussed my projects. Some of the questions which were asked were:

(1) How Logistic Regression works

(2) Significance of A/B testing in product development

(3) How to design controlled experiments for hypothesis testing

(4) Advantages and shortcomings of boosting methods

(5) Parametric vs non-parametric methods of machine learning algorithms

 

Round 4 | Google Meet Interview with client – I

Asked about my WeWork stint in detail. Then proceeded to ask some questions on time series forecasting like:

(1) Given data and the goal is to predict if customers churn two months in advance, how would you prepare the labels?

(2) If customers’ data is in both static and dynamic/time-varying formats, how would you prepare the features and labels?

(3) Before applying any algorithm, what EDA steps would you do and why?

(4) Which algorithm will you start with and how will you go further?

(5) Asked some questions related to ARIMA

 

Round 5 | Google Meet Interview with client – II

Was a fitment round where the interviewer got to know me as a person.

Final Outcome : SELECT

 

What I think worked for me :

Getting the first question of the client round correct. Because as it turned out, they were working on that problem currently.