Are Right, A Lot

Question 1

Tell me about a time you didn’t have enough data to make the right decision.
OR
Tell me about a decision you had to make without clear data or benchmarks.

Answer 1

data = information

1) Situation: Inefficient Excel Sheets
Task: Find Solution
Actions: Collected data/requirements, talked to customers.
Result: Customer satisfaction.

2) Choice of Classification method.
Situation: Joined BSI appointed project leader, had very limited domain knowledge had used out-of-box TextBlob and Vader for sentiment analysis.
Task: Complete Sentiment Analysis project
Actions: Learned about the whole text-processing pipeline and a bunch of classification algorithms for binary classification on a large number of features, picked SVM because of its good performance on high dimensional spaces, and practical reason: because it allowed “partial fit” on scikit-learn.

Result: Obtained correlation of 0.67, compared to 0.3 with other out-of-the-box models previously used.

Question 2

Tell me about a time when you made a difficult decision with input from many different sources (customers, stakeholders, partner teams, and so on).
OR
Tell me about a time when you brought different perspectives together to solve a problem.

Answer 2

1) Choice of Database.

2) DSS interviews and surveys.

Question 3

We don’t always make the right decision all the time. Tell me about a time you made a bad decision.
OR
Tell me about a time when you discovered that your idea was not the best course of action.

Answer 3

3) Times Series project; the same model was done again but with “averaged data”, and i took the wrong average, apologised and thanked for catching the mistake.
4) Search pattern was taking a long time!