Original Author: Huang Chong far, nickname blog worm (male number "Data Hive" ID: blogchong), graduated from Harbin Institute of Technology, five years of big data industry experience, currently in a start-up companies in Shenzhen office in charge of big data, data Hive public number and data Hive studios (www.mite8.com) creator.
In the hustle of big data this business 4 years +, iteration speed of the entire industry, heat and big data technology continues to refresh my knowledge, a comprehensive big data will change our lives, it has become academia, industry and consensus government salary with big data will also be of interest to friends already heard one word "high", really high, of course, provided you have a real technology to support, light force is not bragging use. So how to get into this business and make a reasonable career planning, it became an important issue.
For large data industry status quo and, indeed, some of his views, first, whether right, this stuff is no right or wrong, everyone has their own ideas, so some things, I was willing to share some ideas out speak their minds it ~ ~
Speaking of big data, there is a phrase can describe what it's status quo: everywhere!
Today, in the country, as long as an IT company (say non-traditional industries), we went out, said he feels if the company does not get involved in big data are embarrassed.
So what kind of large enterprises in the data engineers move it?
(1) just washed a year or two, or graduates aspire to contribute to the big data industry.
Big Data industry everywhere just said, the scarcity of personnel, from personal experience, the situation really is.
Business focus gradually shifted to the data portion, the department sharply enrollment (of course there are the old problem of employee turnover), nearly three months, after another interview me about seven people around it. Among the interview two or three years of work experience, there are four or five years of work experience, of course, have just graduated from undergraduate or graduate students.
Look Year feeling pretty good is not it, but if you flip through your resume will cry. To talk about three to five years of experience of it.
CV, the project experienced a big slide management system Han XX, XX electricity supplier back-end development projects, turning eight or nine projects, and finally at the end to see what one or two hot-related big data projects. And to master the technology in a variety of what Spring MVC ah, SSH ah, js ah, even php and the like, only a few of what hadoop what, not daring to put the head, when I started to cry.
This is really the case, foot work experience, many of which are just transferred from other technical fields in which to develop java backend, what MVC framework, such as proficient as the main body of the crowd.
Hadoop can say on how it will point MapReduce, Hive and the like are the norm; will point Spark, write Scala, the Storm know, very few; be able to process the entire data frame it is clear the Wizards ; As for large-scale data mining of the deep, they had this to say "how did the contact, but have the interest to learn."
Quotes indeed the case, a large pit dug too much data, so the individual companies are short of people, but also scarce, and therefore would have I said above phenomenon, various industries, especially practitioners of traditional IT industry, They moved into the Internet, to join the big data. The little big data experience, most of them are fragrant steamed bun was like for the, reluctant to let go.
Therefore, the actual situation is finally on our side, ask HR Editor's Note, said JD HR sent out numerous copies can pull over the interview on Jiang Zi.
The final big boss has spoken, he said: experience almost does not matter, as long as the mind willing to learn to live, we must! So, 7 people, offer to send out four.
But more tragedy yet to come, there are two people probably average 1.5 large data experience, did not get the offer bird, also did not know then gone one company, and the last came two undergraduate and master's graduating students.
So, for now, a hot status quo to bring big data industry is that a lot of java developers change jobs, big industry background data on average more than a year, but despite this, is still in short supply.
(2) Let's look at some "encouraging" recruitment needs.
Casually flip through job recruitment site demand, there are a large number of jobs related to big data is refreshed every day. And then combined with the status quo is some confusion we have just said, you will find a lot of "satisfactory" explanation recruitment.
I hope that the company employing Ye Hao, enterprises or, after reading this, to have a clearer positioning of the hiring.
What we want is big data industry experts!
JD is so described, more than ten years experience in the field of big data, and then will be XX, then they have to be XX. No amount of Anjiu much to say, we have just said combination of big data industry history. ten years? I have a ~ ~ Oh
I have seen this JD, from the most traditional to the IT industry (that did not, the traditional IT industry is also beginning to catch up with the trend), and the Internet company job description to more subtle, at least they will not be at every turn ten in the above "brick house."
But also that individuals want to Tucao is, you said ten years on it, but also to the treatment of extremely odd low. In this regard, the Internet company is more sensible.
About Big Data salary this one, then we analysis ~~
We want to be able to carry out large-scale data mining talent!
About Data mining, also slightly above-mentioned point, the upper application of data mining, this demand with increasingly sophisticated data processing procedures, application data has grown from a simple multi-dimensional statistical analysis, and slowly get to dig deep transition.
Is not that large-scale data, said traditional data mining, in fact, on this domestic situation, is still in relatively early state.
We often see this job description JD, Title written, "Data Mining Engineer", then the proviso that familiar with large data fields will MapReduce, kafka, hadoop, storm, spark, familiar with ETL, NoSQL understanding on a number of familiar , able to build a platform, platform, capable of data processing, will classification, clustering, user portrait, personalized recommendation algorithms.
Written in the last number of years on the job, "1--3 years" (full year is too expensive thing). My God, they are evidently not just want to move data mining engineer ah.
They are like hiring ETL engineers; no, should be big data platform development engineer; nor, indeed seems to move in the data mining engineer, did not see the algorithm needs it.
I catch the foot ah, they are not hiring data mining engineers, they are hiring a versatile engineer, is hiring a God.
So, we see first is the basic ability. Personally feel, basic skills course Needless to say, I prefer to be interested in Big Data technologies, and quick thinking of graduating students.
Why do you say? Because big data technology in this area will involve a lot of new things, a variety of open-source stuff, with little experience does not matter, only the quick thinking enough to have a strong ability to learn quickly, then there is no problem!
Then just the topic, many companies recruit the company name data mining engineers, algorithm engineer, I wonder if they really need the algorithm engineer it?
The answer is clearly no!
I have questions about the difference between data mining algorithm engineers and engineers, discussed with many people, my personal view is that the scope of the algorithm engineer is clearly smaller than the data mining engineer.
Data mining engineers need to understand the entire process of incoming data, including data access, pre-processing, then need to know how to use data to solve real business problems, saying that white is to find ways to make the data generated value.
He needs to know a whole mechanism of data to the business or the output of the system, may involve complex algorithms transformation, it may just be a simple transformation rules, or a combination of multiple models of transformation output, etc., he is a more comprehensive and summarized sexual orientation.
The algorithm engineer is not the same, I think their responsibilities more pure, they need to know how the real problem into a mathematical model, and the model is transferred to the extreme, so as to solve the problem. Therefore, a single algorithm engineers work more content, but more specifically, the need for better math skills.
This is why I can not say the reason of Foreign algorithm engineer, and I fear beat, ha ha ~ ~
OK, a little detour. We look back and say that the current most of the companies who are looking for big data, while also looking for data mining algorithms engineer or engineer.
So how do companies or mining companies in this data to locate it? I personally think that most small companies do not have the conditions to find pure algorithm engineers. If a small company says that you need algorithm engineers, more than either gold wayward, either the name of the algorithm recruit engineers guise, will move the point of data mining people.
As for the reason of it, on the one hand this algorithm in the country are scarce resources, so the cost is relatively high; on the other hand is in the actual business operations, advanced algorithms model is difficult to industrialization (so, something on paper from the majority industrial production is very far away, do not be fooled); Furthermore, is in the field of data mining, some very easy primary industrial production algorithms, and even simple rules customization, at this stage have been able to achieve a business goal, and we why the sweat it? !
So, I think companies in this stage, you can demand that such high-volume data processing, and then know how people conduct data of industrial transformation. Because the algorithm engineer difficult to obtain price-conversion you need at this stage.
Including our Big Data within the department is equally true, the algorithm team crowns the "algorithm" in the title, doing the chores of data processing. It takes time to transition!
Of course, if you must keep the algorithm is a group of professional engineers, spicy it, I can only say, you get too much of angel investment, estimated to be I do not know how to spend, reared on the possessor of it.
As previously said, big data in this area, a little bit of hot deformed, leading to the direct cause of this direction is a lack of people, it is also a lot of back-end java developer career change.
Because the lack of people, they will switch to it? Apparently the pull! Everybody is a good person, would like to see "money." Lack of people, who can not find how to do? Improve the treatment, there will be a natural.
OK, finally learning to borrow big words "as long as the right way to go, not afraid distant", good planning your own career, won a small a target, for example, when the first big data engineer ..... Ok, it First it now!