After two wonderful years in Nord stron Data Lab, I got a job in S3 department of Amazon Web Services. I am excited about the new chapter in my life, and I am glad that the time-consuming and torturous interview process is finally over. Interviews usually include one of three preliminary screenings and a full-day on-site interview. These interviews are full of pressure, because I don't know what I will be asked, and the other party usually expects you to show off your intelligence. Data scientists generally don't do this kind of thing (at least not out of context, just showing off their cleverness on the phone by memory).
You need time.
If you are considering changing your job (or entering this industry), the best advice I can give is to start preparing now. You need to give yourself a lot of time to avoid cramming. Take the time to make sure that you can explain the core concepts in your own language. The question of telephone interview is usually like this: "How to explain to an engineer what the P value is?" Suppose you want to explain to a non-statistical engineer, without using technical terms. There is no doubt that you don't want to explain these basic concepts for the first time on such an occasion. In addition, don't underestimate the influence of nervousness on memory, even if you recall something you think you know well. If you are new to this industry, you may need to give yourself more time to prepare unfamiliar concepts.
I also suggest spending more time preparing personal information, that is, your resume and cover letter. There are two views on this issue, one is that it is important and the other is that it is not important. Will the interviewer really read these materials? It's difficult to give a general answer, but when I was working in Nord stron, I attended a lot of interviews. Personally, I value these materials very much. Spelling mistakes are intolerable. A boastful letter of recommendation does not bode well. Poor profile indicates a lack of interest in the position (or respect for readers), while piling up keywords implies that the interviewer asks the candidate when and where he did it. In the wider technical field, people tend to think what is important on GitHub. But most companies, especially big ones, won't look at your GitHub. They will read resumes and cover letters (this may be surprising, but technology is not ruled by elites). Finally, these documents will show you how to present yourself professionally, so they are really important, even if you didn't think so before.
Experience is the best teacher.
It is recommended to do more exercises and analyze your weaknesses. Many people mistakenly think that repeated reading is the most effective learning method, but it is not an effective way to solve probability problems and logical puzzles on the spot (Make it Stick is highly recommended before you start learning). By concentrating on solving practical problems, you will immediately find your weaknesses and determine your learning priorities. Spending time on what you already know is procrastination, besides, you are already busy. Besides, this is a technical field, so you should be prepared to answer questions at the technical level. If possible, I suggest standing in front of the whiteboard and answering practical questions, so as to adapt myself to this writing style and write while talking. You can find many related suggestions and interview questions on Quora.
Learn as much as possible about your future job.
Do you know what an information interview is? I didn't know what it was until my friend used this method! Sometimes the interview process is advancing, but you don't know whether you want the job or not. Then you can slow the other person down and do an information interview to determine whether this is the job you really want. You can also take the time to "spy" on the company and the interviewer. For example, for Amazon's live interview, I took the time to check each interviewer and their background on Linkedin. This will help you guess the questions they will ask. Oh, this person is an engineer, so she may not ask you statistics, but she may ask about the expansion method. Wait a minute. She is a senior management engineer. Maybe she will want to know my leadership and interpersonal skills. Allen Qiansha has many suggestions on interview taboos.
Get resources!
You can expect to ask about the following fields: statistics, machine learning, prediction, algorithms, all the knowledge that computer undergraduates should know, and the scalability and performance related to all the above fields. Oh, by the way, you should also be prepared to program in the language of your choice. Piece of cake, right? !
book
Make an introductory book on probability theory. It doesn't matter which one. I use a typical undergraduate probability theory textbook written by Ross. If you have this book, I recommend doing a self-test in chapters 1~5 to decide whether to spend more time on it. Combinatorial mathematics and basic probability questions are necessary for telephone interview, so you must master them. I also use textbooks written by casella and Berger to review expectations and variances. This book can be said to be the bible of statisticians. Generally speaking, most interviews are simpler than textbooks.
For computer-related content, I generally refer to three books: Exposure of Programming Interview, Interview of Decoding Code, and Programming Pearl. The first book is the most comprehensive of the three books. If there is only enough time to read a book, then read this book. The second book is concise and specific, aiming at interviewing big companies like Amazon, Google and Facebook, but it is not widely used. The version I used is a bit annoying. Illustrations teach you to be "a buddy who is willing to invite you to drink with the interviewer". The buddy of this book was so angry that I finally gave up reading it (I expected to get more valuable content). The third book has nothing to do with the interview. It is a collection of thinking explanations of computing problems and solutions. This book is not only suitable for learning, but also for understanding the calculation process. If you have time, reading this book will be a pleasure.
Yes, it's about Coursera. If you lose your old textbooks and don't want to buy any books, the information on Cousera can satisfy you. I strongly recommend the biostatistics training camp at Johns Hopkins University. This is an effective review of the first-year course of statistics undergraduate. Don't spend too much time watching course videos. Test yourself with test questions and homework, and then watch the video of weak links. You can also take a look at the specialized courses of data science. The same classes as above are organized by the same group of teachers, including exploratory statistics and R programming. Andrew Ng's machine learning must be learning and enjoying. He is good at explaining the motivation behind the method and spends a lot of time training intuition in the course. Intuition is especially beneficial for telephone interviews. In telephone interviews, you may not want to emphasize technical details, but want to prove your familiarity with the field. Because my goal at that time was Amazon, the special courses of cloud computing also benefited me a lot. I am shifting from retail technology to cloud computing, and I want to better understand the problems I will face. In this case, I only watched the video of the course to learn the vocabulary in this area, without going into the technical details. I always look for good courses on Coursera. Please leave me a message if you have any recommended courses!
Coursera's start-stop time system annoys me. Recently, I found that many courses can read the previously archived materials without waiting for the new class to start. This is an important change for me. Go and have a try!
This is what I call dry goods. But I also want to say some cliches.
First of all, stay calm! Being too nervous will make you forget some knowledge. This is a trouble for me, let me do some crazy things, such as writing everything down and posting it on the wall, but it is not recommended for everyone to do so. My recent madness is doing aerobic exercise a few minutes before the telephone interview, which makes me sweat and gasp. In addition, if you live in the city where the target company is located, interview directly with the interviewer. I have full expression, and I can do better than telephone interview.
Don't forget that you are also interviewing each other. Trust your instincts. I once went to a startup company for an information interview. I feel that the other person is arrogant and doesn't listen to me at all, but I am still interested in that job. I tried to follow up, but I had to make an appointment for 10 thousand times every time to get confirmation. It was a terrible experience. If I had trusted my intuition (these people are unreliable), I wouldn't have wasted so much time. Even the most interesting job is not worth spending eight hours a day with people who don't respect you.
Finally, don't compare your experience with others, because you will be misunderstood or disappointed. When I was going through an interview, it happened that some colleagues I knew were also interviewing. At that time, I was surprised and angry after comparing my experience with others. To put it simply, I interviewed for the same position with a junior male colleague in the same week. He was interviewed face to face by a team member, who asked a very basic question about the probability of rolling dice. I interviewed people in different offices by telephone, and they asked me an optimal solution of game theory. It's hard to accept and it's hard not to get angry. Now I interpret this as lack of recruitment experience and immature company. This company doesn't know how to interview my position, and may even hire people I don't want to work with. I don't want to work in such a place.
Finally, you should be as prepared as possible, but don't get upset because your knowledge is defective. Believe in yourself, believe in your impressions. Learn from the failed interview in order to get the next interview.
;