Refused to resume from today's headlines harvest offer, I spent a year

First, the impression of headlines
Second, the interview background
Three, ready to interview
four headlines side of the phone surface (Java + project)
V. headlines dihedral video plane (large data + project)
six headlines three-sided video surface (algorithm + scene title)
seven how written by headlines?
Eight, how to prepare for an interview headlines?
Nine, tips on how to enter a maker of

First, the impression headlines

1. I am the boss Zhang Yiming told fellow Fujian, Longyan City in Shangougou is my friend say, to get out of the US Mission Wang and Zhang Yiming headlines I mustering the desire to eat Longyan specialty of mouse stem.

2. Give more money, also written thief difficult. Participated in the written headlines students know, headlines written test three times can be written, that is, whether you have not participated in three chances. Written on five algorithms questions, two hours last year, the first time I saw algorithm questions on autism for the first time I hung up. The second and very seriously to do, to give up half. A few days later they invited me to do the third time, there is a feeling of being insulted by force.

3. Push rejected within. As my title says, resume rejected. Last March saw a headline in the big brother says may push spring training on known almost, so the concurrent private letter to his resume. Intense competition in the headlines say, I background is not enough, go beyond that after reading my resume. . (excuse me)

Interpolation rejected

Second, the interview background

Since last year the exhibitions abuse algorithm written headlines and the headlines of the predecessor to the headlines lacks goodwill, and felt no access, so six months later did not concern the recruitment headlines, until the occurrence of the following two things.

1. One day, buddy said before a practice into the headlines, just around the corner and I intend to start to engage. Then there is an immature idea: "Even he could get into, I should be right."

2. In the gym just met a number of positions in the vibrato do little leader, I think this is God ah. So give him a free festival on the private lessons, ask him or her to complete interpolation.

Third, preparing for interviews

In fact, from March to start preparing your resume to the interview in April received offer, but also a month's time, but they are based on my accumulated over the past year, ah, exclusive chair countless times overnight study room, library are I was sitting fall into the trap. Avoid mislead you, do not do this heading the party, wrote a year of preparation, the sense Officer.

book list
They are relatively good book list

The following is my last year to set their own study plans:

Jan:
Big Data entry

The month of January in the brush introductory video and "Hadoop The Definitive Guide", and near the end of the period, 5:30 to get up every day, watching two hours of video lessons and review or final exam.

February:
prove safety offer brush again, data structures, JavaSE, JVM, multithreading

① wins the offer do not understand some of the skip, etc. paint the second time next month;

② data structure focuses on the characteristics of various trees, binary trees, some of the characteristics of the red-black tree and linked list;

③JavaSE main difference class object interface can inherit, the Static characteristics of various principles source Hashmap, ArrayList set like frame;

④JVM learning is based on the integration of high-frequency interview questions online, then brush again face questions only went to see Zhou Zhihua's "in-depth understanding of the Java Virtual Machine", JVM JVM is not difficult layout, garbage collection algorithm these theories, but the combination of code and project scenarios. For example, a piece of code, there are a lot of classes, methods, of which there are global variables and static class, this time to run up the project which, when loaded first, what variables stored in which areas, not only to understand the theory, but also with the code to understand the theory . There is a project in the OOM, stack overflow, JVM tuning parameters and so on.

⑤ multi-threaded brush also face questions online and blog, and several multi-threaded cook cook books are, understanding does not come. The first hands several multi-threaded implementation to achieve again, and then analyze the advantages and disadvantages, and then cut into thread-safe and thus a variety of locks. Multithreading learn different JVM, own more hands-on, multi-commissioning it, the console will tell you the answer.

March:
prove safety offer two brushes, database, operating system, network, preparing for interviews

① understand the code do not understand the idea of ​​time wins the offer a second time focused brush first pass.

② database cow brush off network SQL 100 Multiple - Item combat, others are indexing works, pessimistic optimistic lock theory.

③ frequency OS brush face questions. Generate deadlock conditions, solution and the like, resource scheduling algorithm, concurrency and parallelism and the like.

④ network is also brushed questions. Four three-way handshake waved and why, TCP and UDP, the page request process, DNS resolution process, network model, meaning the response code and so on.

⑤ I understand an algorithm or a brush every sort will face questions, so before finishing the interview to see documents and blog on the line, surprisingly easy. This month also continue to send your resume to interview, from the beginning of March are cast side edge preparation, all ready to go to vote does not exist.

March received the offer Jingdong, seeds and other second-hand car second-tier manufacturers.

April:
Big Data projects, handwriting prove safety offer, continue to interview

① March on three sides by the interviewer Ali rookie sectors hate the situation got that resume is not big data projects face large data Kong, the theory will be more what use he intended to mend the project in April.

② at the same time make up the project, not forgetting algorithm, practice whiteboard to write code.

April received Tencent, where to go for some offer, Tencent and where to face the final scene had to face, so one day fly to Beijing the day to fly to Shenzhen. Tencent offer to get the time to discuss with HR can not go to the base of Shanghai or Beijing, HR said no, so no go. Choose to work and live the city has a rigid standard: far from home! Those students studying in the north after graduation has been to run home, and I was been jumping out, perhaps because in Xiamen for four years now, home is "Assassins" ah.

May-December:
internships, see the source code framework, interview, after finishing surface

Kyligence finally chose to go to practice, just to take this opportunity to learn the source Kylin, watched part of the source Hadoop and Spark. Forced back of the interview but also to their own learning, find deficiencies in the interview, but also the way the information compiled some interviews.

Summary:
Be sure to row their own learning plan, we must learn to adjust the plan according to their own learning. Project is a stepping stone to the interview, but also the bulk of the interview, I was out of practice from the second year, so the project can spend less time here, do not learn from scratch. The underlying theory must be, otherwise manufacturers hopeless. Algorithm must be hard just total on what kinds of questions, I just started also very contradictory, think too hard, behind the difficulties also the case. Not on the first pass the second time, or else a third time, until now, to prove safety offer inside there are some I do not understand the subject, but I will not get to the bottom, the overall situation, not to pull the details.

Here Insert Picture Description

Other students when the winter vacation, a man in his own rented a small house next to the school to learn the eve of the New Year; to consolidate the internship project better interview, ran a few times before the internship company ask my colleagues to dinner and asked details of the project; To interview with Ali just hard liver, review the night; cast your resume too, information was sold, and heard the phone ring thought it was a telephone interview; go Tencent end face, in the deep of the night reviewing all night study room.

These are my last year of preparation, tell me your resume ready to receive offer this month.

1. Review algorithms. The question before the brush-off algorithm to sort a bit, because there are collated before, so how much time to review them useless, but also specifically to understand the long KMP algorithm, LRU algorithm.

2. Review the theoretical basis. In addition to data compilation and review by the side of, but also specifically to understand a bit Golang, because the headlines go with the language.

3. specially prepared technical difficulties on several projects, you can dazzle during the interview.

Fourth, the headline on one side (Java + project)

1. inverted index
2. redis talk about inside the hash table?
3.happen-before rules?
4.volatile modifier, synchronize lock
to achieve Singleton pattern 5.java, lazy, hungry man?
6. processes and threads difference, the difference between multi-process and multi-threaded?
7.HashMap principle, why does the red-black tree, features red-black tree?
8. quick drain of time and space complexity, the best worst-case scenario, the optimization program?
9.TCP congestion control, the specific process is kind of how? UDP congestion control it? How to solve?
10. Learn to talk about garbage collection algorithms and collector, when executed STOP THE WORLD?
11. Go to understand language?

Fifth, the headlines two faces (Big Data + project)

Project architecture 1.Kylin of
2.Paxos and ZAB agreement
3.CAP theory, fault tolerance, meaning a partition
4. Join the large table small table optimization, how to handle data skew?
5. talk about maximum and minimum heap heap
read 6.HDFS, written, or fault tolerance. (Source)
process 7.MapReduce the (first and second editions of)
8.MR shuffle, shuffle the Spark.
9.namenode HA, split brain, Yarn scheduling mechanism.
Inner and outer tables 10. Hive difference, the number of bins modeling model, the number of hierarchical positions, snow model and star model.
11. ClickHouse understand it? It differs from the Kylin?

Sixth, the headlines on three sides (+ algorithm scene title)

1.LRU algorithm (pseudo-code)
2. K chain penultimate number (say ideas)
3. pile of screws and nuts matching the shortest possible time (code implementation)
4. seek new user browses pages per day (Hive QL achieve)
the maximum vibrato small demand video clicks per day 10 (Hash + minimum heap)

Seven, how headlines written?

The written test in a total of five, I AC 2.5 title, the first title of 90%, 90% the second question, the third question 70%, not good, but barely passed. This time last year but even I can not read the title, can understand his mouth only know how to say, not implemented the code. By this time the training algorithm, summed up the following points algorithm written Experience:

1. usually brush topic must be summarized, the best classification. For example, questions about trees, linked lists, arrays, and so on, observe their problem-solving ideas, summed up the problem-solving routine.

2. accumulation tools algorithms. What tools algorithms? Algorithm is solving a problem you need to use another algorithm, this algorithm is called the tool is to solve this problem of road algorithm. Such as the common "depth-first traversal," "breadth-first traversal" Tools algorithm "01 backpack", "KMP algorithm" as well as common selection and sorting algorithms are frequently used.

3. Learn abstract topics. Algorithm is different from the written interview questions algorithms will not tell you what kind of straightforward algorithm to be used to answer more to learn to abstract themselves, put aside the topic itself, to understand what is inside talking about, do not be the subject of sugar-coated bullets confused. Only the title of the abstract into the most primitive algorithm you can make better use of tools like algorithms to answer.

Eight, how to prepare for an interview headlines?

Headlines interviews and other manufacturers may be similar, but this is a little different is the key place where people can be different from other places. We are prepared to do the following aspects:

1. specially prepared several interview questions source. I like to answer some law-abiding face questions that they will depart from the source point of view. Such as the common source portion HashMap other collections, as well as a variety of multi-thread lock large data frame.

2. specially prepared project difficulties. Some manufacturers face when asked about the project can not avoid the difficulties, it is impossible skated over ordinary Bug ah. From the technical side of the stack can be used in the project, to find and difficult technology stack in the project will exist, then set into their own projects, find themselves can understand, it is best to understand the whole.

3. targeted to understand the company's technology stack usage. For example, use headlines Go, well-known, why not surprise me to learn about it? For example, the industry learned from the first use of headlines is Kylin framework, slowly back into ClickHouse, then I have to look at the difference between the two ah, but also can be combined with headline business scenarios.

Nine, a recommendation on how to feed small manufacturers
I have been advocating "Quxianjiuguo" approach, this approach has also been spread to others. Now lack the ability to background does not matter, as long as planning their own learning path, climb a ladder step by step, you can always achieve your goals. Because two of my own background, but I realized early, sophomore began to practice, with project and internship compensate for the lack background, then go step by step "offer freedom." Many times people can not only huge success, you only need to do the ordinary people on the move, you can be more than a lot of intellectual giant, dwarf action.

Released five original articles · won praise 128 · Views 6746

Guess you like

Origin blog.csdn.net/BigData98/article/details/104824162