Xerox Budding Scientist Interview Questions and Answers

Round: Technical Interview
Experience: I got stuck in the question as I wasn't able to approach this question. The interviewer helped me out by giving me the following two clues.
1. Two properties of MST : Cycle Property and Min-Cut Property.
2. 4 cases : (modified edge in old MST)*(weight increased or decreased) -> I eliminated two cases as it would return the old MST. Only 10 and 01 need to be considered.
Tips: Requires Graph Theory(Min Spanning Tree algorithm, Min Spanning Tree properties etc.)

Round: Technical Interview
Experience: I had no idea to approach this problem.
I decided to first come up with an algorithm to obtain the maximal transitive sub-tournament given a tournament.
The interviewer then told me to see if I can upper bound the minimum out-degree in a tournament.
After this, the solution was to upper bound the minimum size of the transitive sub-tournament.
HINT : Use the fact that a tournament contains N_choose_2 number of edges for part II.
Tips: Requires very little graph theory. More thinking based, proof oriented. I had to prove every claim I made.

Skills: Graph Theory, Logical And Structured Thinking, Proving Abilitites
College Name: IIT Madras

Round: Test
Experience: There were questions based on probability, algorithms and graphs. Probability questions were simple. Algorithms portion was mainly on different sorting algorithms and their runtime. Graphs on graph asked about complexity, number of graphs possible, etc. Based on the scores, a shortlist was prepared for campus interviews.
Tips: Solving puzzles from -----/ and "Fifty Challenging Problems In Probability With Solutions" is good enough preparation for any question asked on probability during placement tests. For algorithms, a data structure course done in my college was helpful.
Duration: 30 minutes
Total Questions: 20

Round: Technical Interview
Experience: I was asked to explain my projects. After that, I was asked questions on PCA, SVM, KNN, linear regression, logistic regression, entropy, information gain, decision trees. The interviewer asked for equations in all topics.
Tips: Remember details about every aspect of your project. Even minor details like which features you extracted, why you chose certain classifier, etc were asked.

Round: Technical Interview
Experience: The interviewer asked me to explain about something innovative I had done in my projects. Then he asked a guesstimate question - Estimate the number of trees and plants in a city. This was followed by another guesstimate - How would you estimate the rate of growth of trees and plants if you have data only from the present. (no past or future statistics can be used)
Tips: Practice guesstimates as done for case interviews. Practice presenting your projects in an interesting manner with your friends. Ask your friends to ask you questions about your projects. Since they have heard about your project for the first time, they may ask you questions which you never thought of. This should be sufficient to get past the interviews for Xerox Research - Budding Scientist profile.

Skills: Algorithms And DataStructures, Guesstimate Skills, Machine Learning
College Name: IIT Kharagpur