1. 1st round involves question related to projects and earlier experience. 2. 2nd round is to develop model to classify for a given geolocation the position has rooftop or not. 3. Theoretical question simple based on performance metrics like F1 score, precision, recall, MAD which were used in my projects.
The assignment was a real estate price prediction problem ,they ask you to provide insights and do Machine learning to predict the price of units in Greece . Here when I really became disappointed , the data science consultants ( or who claim they are) who interviewed me lacked basic intuitive knowledge about data science , questioned the why behind creating a simple benchmark to validate ML models on . They also seem that they didnt try to solve the problem before they gave out to candidates , which made them unaware of the aspects of the problem
Standard questions for such positions! SQL, Statistics, AB-design, Time Series, Data Engineering
Previous experience in similar projects
- ML Modelling problem - Talk about a recent data science project you have worked on
No technical or hard questions. They asked me to talk about the major projects that I've been working on.
How would you calculate the cost of a missed shipment?
What is funneling method in marketing.
sql query for joining 3 tables with duplicates and cte with some interval chaining logic
They asked a combination of problem solving, programming, CV, ML, DL, DS questions. Due to NDA, I cannot reveal the questions per se, but here is an overview of some concepts they tested. CV questions- (Resume related and some applied problems) CNNs out there, basics, losses, etc... Favorite papers and details of architecture GANs Programming: Standard easy-medium leet code questions Sampling, probability, python programming. Other ML/DL questions: Ensemble methods Classification basics - loss probability basics recall, precision, ROC etc clustering basics + advanced dimensionality reduction Interviewers touched upon almost all CV, ML and DL basics. They asked a bunch of applied questions too.
Viewing 2821 - 2830 interview questions