Many graduates want to engage in big data development, but there are not too many complete ashore process on the Internet. Most of them are interviews and so on. So I want to sum up, I don’t know whether it can help people in need.

This article is from the core technology group

Personal background

I feel it is necessary to write personal background, without personal background of the ashore process is to play rascal, some people said at the beginning without competition, without a paper, I think that I will learn, finally know is Fudan this master, this you are not to make my state of mind yao. (Dog head, if you offend, don’t hit me.)

I am a student of big data for one year. I like rap basketball.

Ashore process

Rely on own love study, get big factory summer internship, then autumn recruit ashore. A simple sentence requires elaboration.

First of all, you need to love big data. Love is the first driving force. I know many graduates choose big data because they think that the back end or other things are too big, and choosing big data is a way to escape.

  • There are far fewer jobs than on the back end
  • Your practical work experience is really needed
  • My study and work have been seriously derailed
  • Many large factory posts are extremely detailed, you may see the data development is not suitable for you

I personally feel in the big factory fresh students and volume of the situation, the big data gradually fairy fight, if it is a first-tier city, that is sorry, the gods down to earth, that as a double non how to land?

  • First, whether you really like big data

The harder it is contradictory, although say the love, but after a start to learn whether like big data to you for your help really very big, can you accept from 0 begin to contact a lot of components, and may be your undergraduate course completely haven’t heard, if after contact with love, so congratulations you, you have the same starting line and others, but running speed can be fast, Learn what you love every day, is how happy things, can let you finally stick to it. If not really like, just want to find a job, volume back-end/front-end/client rely on more reliable, since all volumes, choose more posts, easy to enter.

  • Second, lay a solid foundation

Note, I am here refers to is not the basis of the basis of large data, but your computer’s base, popular said is regular literacy training, more specific points is one’s deceased father grind, 408 subjects (+ + + computer network operating system data structure) of computer constitute principle, why like by training, is because he did not how to learn, at least one DiuDiuDe impression in my mind, I benefit a lot from the foundation these subjects have laid for me, which is helpful to my way of thinking and even to solving problems. Besides, my ability to read computer papers, read documents and draw pictures are all cultivated at ordinary times.

That if not by training, or really learn bad what can we do at ordinary times, that is, of course, to make up for, but if engaged in large data is determined, I feel enough + operating system data structure, can be a lazy slightly, the importance of these two subjects is affecting me all the time, his thinking on the computer a lot of areas have the shadow.

The rest of the basics, programming language, big data ideas, what frameworks you learned, and so on. The biggest difference between school enrollment and social enrollment is that the most important thing of school enrollment is to show their potential worth training, you learn everything in the interview is to show this point, I give an example of the interview encountered a question: C in the garbage processing fast and Java itself by JVM processing, do you think there is a difference. This problem is the interviewer combined with background and chat temporary asked before, if you are full of backrest stereotype directly cool, need you think answer on the spot, he don’t need you to answer the standard answer, even the problem, there is no answer, but the interviewer will pass your thinking to determine your ability to think on the spot, how to think and answer, the answer is a, the foundation is solid.

So, if you still have a few years to go, lay the groundwork

  • Guess what? Autumn recruiting actually started in March

    Double into giant old hard now, best to find a summer internship, continuously is big company, but need to find a, basically giant summer started in March, and some is the autumn, the default, you can become a full member, I had an interview with just a few companies and I said, this is the autumn recruit direct positive, so can ask a little bit difficult, Qiu Zhao will not release HC.

Having internship experience helps me a lot. The main gains are as follows:

1) Engineering ability, github/ Maven /shell/ Linux, these common engineering modules are rarely used, and I am familiar with some test handover processes to know how the work is in reality.

2) The end of the actual massive data, usually practice hundreds of thousands of pieces of data at most, the cluster on the machine is also very few, internship let you actually feel billions, tens of billions of data volume, and the huge cluster, harvest a lot, will encounter many unexpected problems, can be a thorough introduction to big data.

3) of autumn recruit a great help to you, did not say company internship if confirmation, that you have a lot of autumn recruit them, the key is if you don’t become a full member went out looking for work, others see, internship, it means that you have some actual work experience, has the potential to companies believe that at the same time, it also let me have the confidence to oneself can believe in your potential.

So, for the big data of school enrollment, there are not only big data, but also many other things, but ultimately it is a process of showing their potential.

What to learn

  • Programming language classes: Java, Scala (Fur), Python (fur),shell (fur)

In fact, it is mainly Java. In the interview, I also encountered Java, including concurrent programming, JVM GC and basic collection

  • Data structure ！！！！ : Force buckle does not brush to hang directly, unless educational background is very good

  • Basics: Take a look at data structures and operating systems if possible

  • Hadoop components: Hadoop (basic Mr Job, see Google troika), hbase (basic application), Hive (HQL)

  • Spark: Sparkcore, SparkSQL, sparkSql, and spark Streaming

  • Flink: If you have enough time, you can read it (recommended introductory books flink basic tutorials, as well as the official flink documents), and then do a simple project or get started quickly

  • Message queue: Kafka, learn about it and use it, even in your own projects

  • Tools brought to the project: Flume, SQoop,datax, Redis. To be honest, I know I have used it after the project, but if I use it again, I will still write against the document

  • Projects: Visual small project (Java SSM framework basic practical), offline data warehouse (mainly Hive), real-time data warehouse (Flink simple practical)

The rest is something else that I learned during my internship, so I’m not going to write it.

What you come across in the interview

Interesting scenarios or open questions

1) How to design an arrayList with no storage limit

2) How do you see the difference between C and JVM when it is faster to recycle heap

3) Two super large tables do association, if not SQL, let you directly design the calculation process how to design

4) Conventional million-level data blablabla…

5) How do you research a framework

6) Java has unit tests, how to design SQL unit tests

7) After actually updating SPARk3, are all the advantages compared to SPARK2 in the actual work

8) How to determine the occurrence of shuffle and how to quickly locate the problem

I can’t remember all the questions, so I chose some interesting ones. I didn’t give my answer at that time. Everyone has his own way of thinking, and I hope your thinking can be divergent.

At the end

The readers of this article put together the problems encountered in spring recruitment/autumn recruitment, and make a simple summary. It has certain guiding significance for those who want to transform big data development. You can also go to Niuke to see other people’s face books. If you really love big data, please stick with it. It’s still fun.

By the way, if you want to join my core technology exchange group, like the excellent group friends who submitted this article, to learn and communicate with more excellent partners, you are welcome to scan the code and add my wechat zwj_bigdataer, remember to note [nickname – city – post], I will pass the first time after seeing it.

👇 Welcome to follow the wechat public account: “Big data Dreamer” 🔥 :

A big data enthusiast who likes reading, output and replicating. Besides sharing basic principles of big data, technical practice, architecture design and prototype realization, I also like to output some interesting and practical programming content and reading experience……