4 ms·

I have a series of basic questions I include in any data science interview, and one is "please describe what the central limit theorem says in simple, high-leve

by rm999 12y ago

I have a series of basic questions I include in any data science interview, and one is "please describe what the central limit theorem says in simple, high-level terms". It's absolutely amazing how many people who have great credentials can't do this. I get a lot of "any distribution becomes normal when you sample it enough". This is nonsensical and shows a lack of understanding of the theorem. Please, if you claim to know stats, understand what the central limit theorem says. It's a pretty incredible and useful theorem.

whorleater 12y ago
Out of curiosity, mind sharing some of those questions? I'm a computer science major looking into getting a data science internship, and I'd love for a chance to see some real interview questions.
- rm999 12y ago
  Yep! Here are a few, I customize the questions a bit based on seniority and background. 1. In a 1-d gaussian with mean=0 and variance=1, what is p(x=0)? 2. If I give you a csv file with a header, how would you return a tsv file with columns 6 and 2? Just high-level, not actual implementation. 3. If I have some dataset (I change it up a lot, a common one is subway ridership by station and date), how would you create a visualization to give some insight to a fairly non-technical audience, e.g. the distribution of ridership by station. 4. If I gave you a house dataset with a bunch of attributes (e.g. year built, number of sq feet, presence of pool, zip code), how would you go about building a machine learning model that predicts whether the house is being rented or is owned (assuming you are provided with tagged data)? The selling price? How many times it has been sold? 5. What's a technology that you're excited about (could be a computing tool, a machine learning algorithm in a library, etc)? Tell me some of the pros and cons of it and competing technologies. We also give out a 4-6 hour task as homework, and evaluate the code quality, tools used, answers provided, etc.

by rm999 12y ago

whorleater 12y ago

Out of curiosity, mind sharing some of those questions? I'm a computer science major looking into getting a data science internship, and I'd love for a chance to see some real interview questions.

rm999 12y ago
Yep! Here are a few, I customize the questions a bit based on seniority and background. 1. In a 1-d gaussian with mean=0 and variance=1, what is p(x=0)? 2. If I give you a csv file with a header, how would you return a tsv file with columns 6 and 2? Just high-level, not actual implementation. 3. If I have some dataset (I change it up a lot, a common one is subway ridership by station and date), how would you create a visualization to give some insight to a fairly non-technical audience, e.g. the distribution of ridership by station. 4. If I gave you a house dataset with a bunch of attributes (e.g. year built, number of sq feet, presence of pool, zip code), how would you go about building a machine learning model that predicts whether the house is being rented or is owned (assuming you are provided with tagged data)? The selling price? How many times it has been sold? 5. What's a technology that you're excited about (could be a computing tool, a machine learning algorithm in a library, etc)? Tell me some of the pros and cons of it and competing technologies. We also give out a 4-6 hour task as homework, and evaluate the code quality, tools used, answers provided, etc.