Interview Journal (Internship)

Posted on 05/02/2017, in Career.

2015.10.20 9:00am - Writing test - Credit Suisse

  • Feedback: C++ class, inherent; discrete probability; stochastic process

2015.10.21 9:30am - Interview 1st round- J.P.Morgan

Quant research under corporate finance 1.ADT(Abstract Data Type); 2.implement calculating rolling VAR;

  • Feedback: Xueqiu project; P or Q quant

2015.11.9 8:30am - Interview 1st round-Goldman Sachs

Tech: 1.Finance-Call spread: Payoff; Delta; Where Delta is at its max? Is the gamma graph symmetric? 2.Programming-Binominal tree: Longest path in the Binomial tree; Computation cost (best case and worst case); Depth-first search 2.Programming-Binominal tree: Shortest path in the Binomial tree; Computation cost (best case and worst case); Breadth-first search Feedback: Basic data structure, more intuition on derivatives, more practical questions.

2015.11.12 11:00am - Interview at Xtend, 1st round, SAP

Software company, silicon valley innovation center Tech: 1.What does rank of matrix means? 2.Smater pointer 3.Public inherence and private inherence 4.Resume: K-means, parameter in random forests Feedback: Prepare details of the machine learning method mentioned in the resume.

2015.11.12 11:30am - Interview at Xtend, 1st round, BNY Mellon

Bank, Asset management team and innovation center Behavioral: How to get along with difficult people

Tech: A mechanism deal with forgotten password

2015.11.12 4:30pm - Interview at Xtend, 1st round, infosys

Data driven consulting, indian company Lots about introduction No tech no behavioral

2015.12.3 1:20pm - Morgan Stanley, Global Capital Market Team, Henry Fan

Henry Fan -> Stanford MCF 13’ 1.Quantitative Finance a) What is delta and gamma. How to do a delta hedge? What if Gamma is large? b) S = 120, K =110, v = 0.4, T-t = 1/4, delta(c) = 0.4. What is the problem? c) Let’s say in the previous problem, delta(c) = 0.6. What is delta(p)? How can you justify your answer.

2.C++ Feature a) What is polymorphism and inheritance ? Can you give a real-life example that polymorphism is useful? How to implement this in C++(virtual function)? b) Is it possible to implement a super class of call options with the methods of pricing that has polymorphism feature on two sub-class American Call and European Call? c) In b), if Monte Carlo is used, what’s the trick?

3.Algorithm a) Find 3rd smallest value in an n-entry array in O(n) b) Find k-th smallest value in an n-entry array in O(nlog(n))? c) What is the best case and worst case if quick sort or merge sort is used?

4.Advaced Machine Learning Algorithm a) What is SVM(Support Vector Machine)? Can you explain this algorithm and why is the support vector crucial?

Take-away: a) Taking MATH238 and CS229 next quarter. b) More leetcode in array, linklist and DP. c) C++ features.

2015.12.9 9:30 am - Goldman Sachs super day 1st

  1. Resume ===========TECH============
  2. Statistics - Maximum likelihood: 2.1.If you found some distribution roughly like Exponential distribution but it’s discrete, how would you like assume its PDF? 2.2.Is there any problem in the PDF of parametric method? How would you adjust it? 2.3.Estimate the parameter using MLE.
  3. Coding - Dynamic programming (also explain DFS, BFS and DP): 10 choices at 10 periods; each have different pay-off; if you switch your choice, you need to pay a switch cost according to a switching matrix, finding the path that gives a maximum pay-off.

2015.12.9 11:00 am - Goldman Sachs super day 2nd


  1. Quantitative Finance - Interest rate swap: 1.1.An interest rate swap, one side pay the corporate 5% rate, the other side pay the interest rate at each period, how to price this swap? 1.2.How to hedge the corresponding corporate coupon? 1.3.What is interest rate goes to 0 or Inf?
  2. Quantitative Finance - Foreign exchange rate swap: 2.1. Suppose you have a pen company in Europe which sells most of its products to the U.S. What will happen to your profit if the EUR/USD goes up? 2.2. Suppose the EUR interest rate is 0 but the USD interest rate is 5%, which currency will be weaker in one year? ===========BEHAVIORAL===========
  3. Calculate 15*16
  4. What will you do except financial service industry?

2015.12.9 4:00 pm - Goldman Sachs super day final

  1. Resume: 1.1. Transformation function of Logit/Probit model 1.2. How to estimate Delta? 1.3. How to calibrate stock position more efficient than brute force?

  2. Coding - Dynamic programming: House robber problem (see LeetCode)
  3. Brainteaser - Induction: Suppose there are n thread in a bowl (2n ends). You randomly pick up 2 ends and connect them by a node until there is no ends in the bowl. What’s the expectation of loops in the bowl.

2016.1.12 2:00 pm - LinkedIn 1st round 1.1. Construct a tree from the dfs traversal 1.2. How to estimate the seniority from the LinkedIn website

Feedback: Even in a co-editor, the most important thing is algorithm

2016.1.15 11:00 am - Citadel 1st round, on campus 1.1.Baysians law to detect true alpha holder 1.2.Deck: 1 for black, -1 for red, stop as wishes 1.2.1.Write the recursive expression 1.2.2.Write it in DP 1.2.3.Optimize memory O(min(m,n)) 1.3.Maximum likelihood 1.3.1.Estimate the p of a coin if p is uniformly distributed 1.3.2.Estimate the p of a coin if p is of other distribution

2016.2.20 - 12:00 am

Akuna 1st round, video 1.Braintesaser: three bags (apple, orange, apple and orange) with all the tags wrong. How could we make them right by one pick? 2.Python: How to remove duplicated elements by using the build-in data container? 3.Python: What’s the difference between tuple and list?

Takeaway: Prepare for language features of C++, R, Python. At least for the basic data container.

2016.2.25 - 11:00 am Bridgewater, group interview:

1.Is it fair to use PED to improve the athletes’ performance? 2.Is there some particular policy should be implement if we want to regularize the market of human organs?

Takeaway: Do not apply for those positions which requires heavy verbal skills. Waste of time.

2016.2.26 1:00 pm - Viking Global 1st round, phone:

  1. Statistical hypothesis testing: 1.1.How to compare two strategies’ performance? 1.2.How to compare multi-strategies’ performance? 1.3.What’s the problem of using pair-wise t-test rather than F-test? 2.R: Explain the following functions: 2.1.with 2.3.aggregate 3.OOP: Explain the following concept: 3.1.Encapsulation 3.2.Method overloading

2016.1.28 1:30 pm - Viking Global 2nd round, phone:

1.How to conduct a hypothesis test? 2.Explain the concept of PCA 3.The basic assumption of linear regression 3.1. What if the error is not normal distributed?

2016.2.3 10:00 am - Viking Global Super day:

======= 1 ========== 1.STATS: What’s the assumption of liner regression 2.STATS: How to prove the liner regression estimator is the best liner estimator? 3.STATS: How to generate random variables of correlation p? 4.STATS: What’s the property of correlation matrix? 5.STATS: What’s the range of pair-wise correlation p for n variables? ======= 2 ========== 1.Behavior: What’s the difference between America and China? ======= 3 ========== 1.Trading: Name several factor that affect the trading cost. 2.Trading: How the volatility affects the trading cost? 3.Trading: Executing in the morning, noon, afternoon, which one will have the largest trading cost? Why? ======= 4 ========== 1.Database: SQL subset data query 2.ALG: find duplicate in an array 3.ALG: Merge sort/Quick sort 4.ALG: How to implement a hash map ======= 5 ========== 1.Numerical: Design an ALG to calculate square root and prove its convergent rate. 2.Numerical: If the SVD could be used to get eigenvalue?

2016.2.9 10:00 am Facebook, 1st round, phone:

1.STATS: Probability of a pair of same number/color cards from random draw 2.Database: summarize table: story-comment data 3.Business sense: How to detect the advertisement/noisy comment under a story

2016.2.20 7:00 pm - J.P.Morgan, Beijing HQ, 1st round, phone:

1.ALG: 3sum 2.Stocastic process: definition of brownie process 3.Brainteaser: If an unit stick is broken in two place randomly, what is the probability that the three pieces could make a triangular? 3.1. What if firstly the stick is broken into two, and we randomly broken the larger one?

2016.2.22 3:00 pm - Facebook, 2nd round, phone:

1.Database: Summarize table: SMS-country-carrier-time 2.Business sense: What’s the distribution of high school count in Facebook 3.Business sense: How to detect the fake high school information/typo in the Facebook Profile, based on location information.

2016.2.26 1:00 pm - Amazon, SDE, 1st round, phone:

1.ALG: Group Anagrams 2.OOP: Design a deck 2.1. Implement shuffle method using a random number generator

2016.3.1 4:00 pm - TuSimple, phone:

1.STATS: Why is statistical test based on p-value of Gaussian distribution not suitable for bankruptcy analysis? 2.CS: Which part will be the bottleneck if we want to iterator through millions of picture (with their meta data)? 3.STATS: Which statistics could depict the concentration of a discrete random variable? (Entropy) 4.STATS: How to estimate hidden parameter?

2016.3.3 11:00 pm - J.P.Morgan, Beijing HQ, 2nd round, phone:

1.STATS: correlation matrix, range of correlation 2.Stocastic Process: E(W_t|W_1=8), 0<t<1 3.MATH: e^\pi, \pi^e, which one is larger? 4.FINA: Two proflo, both are made of two vanilla European call options on A and B. The first one has strike price at S_A and S_B respectively. The second one has strike price both at \frac{S_A+ S_B}{2}. Which one should be cheaper?