Statistics interview (AMAZON) 1. If there is stream of data x1,x2,x3, …. Having multiple Gaussian distributions. How will you allocate point A to a particular Gaussian distribution. 2. In what conditions K-Means don’t work 3. If there are 3 recommendation algorithms, that generate recommendations for Amazon Search bar, how will you select the one which is the best? (Use precision of different ranks and combine into 1 metric) 4. How will you do anomaly detection ? 5. How will you calculate bid price for a DSP if we want to generate 2 times ROI? 6. How will you evaluate CTR Predictor offline and online? 7. How will you handle imbalance data? What are the different metrics that can be used? 8. If we have two time series data: One is hourly time series sales values, other is overall sales value (like amazon keeps on adding new products and sale is increasing linearly). How will you find if there is anomaly sales at particular time? 9. If you have observations, that can be separated using a line..what will be the equation of that line SQL and Python 1. Left join, right join, self join, cartesian join 2. Lambda function 3. Pandas data manipulation 4. Find max salary of each department 5. find average salary based on duration