[Interview Experience] HCL | Data Scientist | Round-1 | Cleared
Ankit Thummar
874
Machine LearningInterview
  • YOE: 12 Years
  • Relevant Experience: 4 Years (Experience on tech position: 0)
  • Transitioning from chemical manufacturing to data science

I applied for the position of Data Scientist at HCL Tech through Naukri.com and got a call for first round interview within 3 days. After rescheduling it one time by HCL I finally appeared in interview aiming to transition my professional career into Data Science field as I didn't work with any tech company professionally.

As interview process was online via MS Teams, at the beginning of interview they asked to show my ID proof and captured screenshot of me holding my ID Proof. Then begun with basic introduction of education, work history and later explained a machine learning project I have worked on.

After this started with technical discussion. First thing I was asked was about how to reduce dimentionality when we have 4000 columns/features and 1 target variable (discreet or continous). I explained how can we eliminate multicolinearity, do feature selection and perform PCA.

Then I was asked about which machine learning model I know best, to which I replied Timeseries Forecasiting, Recommendation Systems. Then we discussed in depth about timeseries forecasting. What is non-stationarity? How can we detect and remove it? How to find values of AR(p), MA(q) and other parameters? etc.

For superwised algorithm I mentioned my best knowledge abour Naive Bayes, and was given two different scenario. In one all features are numerical, another all features are categorical, for which case I should go with Naive Bayes and why? To which I replied, for numerical assuming data is normally distriuted we can go with GaussianNB and for categorical we can go with BernoulliNB (I should've replied CategoricalNB instead).

Later we moved on to discuss hyperparameter tuning method and discussed about GridSearchCV, how it works, what are the consideration we need to take care etc.

Then we moved on realtime coding challange. I shared my screen and VSCode window to answer the coding question.

  • Challange: You are given a comma separated text file where first row is column title and remaining rows are numbers seperated by commas. I need to perform row wise sum of all the difits without using any standard library.
  • My Code:
with open("data.txt") as file:
  data = file.readlines()
for i in range(1, len(data)):
  nums = data[i].split(",")
  total = 0
  for n in nums:
  	total += int(n)
  print(f'Total sum for row {i} is: {total}')

Finally, I was asked about experience in GenAI and LLMs, to wich I replied in brief about authentication process about openai library and how to use it to train LLM with good prompting skills.

Ultimately I was advised to focus more in GenAI if selected for next interview round as proposed position specifically demands majority of GenAI work.

Overall it was great interview experience .

Comments (3)