►Editor's note:
Do many students think that they can sit back and relax for the time being after getting a graduate offer? It would be a mistake to think so, and using the summer vacation after your senior year to prepare in advance will not only allow you to fill in some of the blind spots in your job search knowledge, but also avoid being in a hurry when school starts. Today, Tutor Sophie will take you to understand the job search strategy of data graduate students~
►Overview of the content of this article:
☑️ Different DS job types
☑️ Preference of various technology giants
☑️DS Interview Process
☑️ Admission-graduation job search timeline
☑️ Resource Recommendation & Preparation Method
☑️ Interview preparation advice
► Different DS Types:
There are four main types of data scientist jobs that are commonly found in the market, and different types of positions will correspond to different skills and interview content.
Data Science Analytics（DSA）
Mainly data analysis, A/B testing, and communication with other cross-functional teams such as Product Manager/Software Engineer. In the interview for a DSA position, in addition to the technical skills of SQL and Python to analyze data, the ability to communicate and interpret data results is also examined.
Machine Learning Engineer（MLE）
This type of role is more focused on modeling and machine learning skills, and usually requires a very deep understanding of machine learning and the ability to turn a research-stage model into a production-ready product. Therefore, in the interview for MLE positions, there will be more in-depth investigation of machine learning theory and model development ability, and there is less high requirement for communication or ab test ability.
Research Scientist
Many big companies like Meta and Apple deploy jobs like this, which are actually very similar to software engineers, requiring a lot of coding and focusing on model development. For example, when I worked at Meta, the research scientist in the group was actually the boss of the report to the coder, so it is actually equivalent to a coder. These positions usually require a relatively high academic background, such as a PhD.
Data Science Consultant
This type of role, which is often found in large consulting firms such as McKinsey and BCG, requires a combination of DSA and MLE skills, a combination of SQL analysis data, and a background in machine learning.
More industry-related introductions:Data Science & Data Analytics Industry Introduction (Part I)Data Science & Data Analytics Industry Introduction (Part II)
►Preferences of various technology manufacturers:
Next, let's briefly introduce the preferences of various technology companies (for reference only, or the actual job description shall prevail).
· Meta比较倾向于招Data Science Analytics和Research Scientist。
· Google is more inclined to hire Product Analysts, and the job requirements are more similar to DSA; You may also often see that Google has several different DS job titles, which are closer to the aforementioned MLE.
· Netflix is more MLE-oriented, and I rarely see them hiring for analytics-related jobs, and the company prefers candidates with years of experience.
· Apple的话对于Data Scientist和MLE也有较大需求,po出来的DS岗位一般是要求有较强machine learning背景。
· Airbnb also recruits a lot of DSA and MLE.
In general, different types of DS roles have different skill requirements and interview content, so knowing the type of role you're applying for and the preferences of your target company can help you better prepare for the interview.
►DS Interview Process:
In the case of DSA, there are usually three parts to the interview:
1) HR Screening first talk to HR to make sure that your background is suitable for the position, and he will ask you about your visa status, when you will graduate, etc.
2) Technical Interview examines basic coding and product sense skills. Among the dozens of companies I have interviewed, SQL is a mandatory item; I also know that the SQL courses in many schools do not match the content of the interview, so it is important to practice the SQL skills thoroughly in your spare time. Python is not required, but you still need to know some. Interview questions are usually broad and require demonstrating good communication skills and the ability to resolve ambiguous questions.
3) The final interview is often a full day of back-to-back interviews, and the interviewer will be more senior than the previous one, which may be the VP manager or even the big boss in the group. The following aspects may be examined: first, BQ and Team Fit, whether they can fit in with the group's culture and whether they can cooperate smoothly with the people they serve after work; The next common test is Product Sense, which will be more relevant to the actual content of future work; Finally, there is an examination of the basics of modeling, machine learning, and statistics.
► Job search timeline from enrollment to graduation:
The sooner you start preparing for a job, the better, and the more efficient it is to prepare when you have a clear direction.
After reading the above content, I believe you have a certain understanding of the job types of DS and how to distinguish which type of position a job description belongs to. At this time, students can search for some companies they like with new knowledge and write a list to divide the companies into different tiers, such as:
最想去的dream company
The company which I want to go to in the middle
You may not go to a company where you get an offer, but you can use it to practice your interview skills
During the preparation period from April to August, you can start to network these companies in a targeted manner, and talk to the employees who like the company to establish connections. When the post opened in August, I "pushed the boat down the river" and asked them to push internally. This is also a time when you can start learning the skills you need for interviews and lay a good foundation in advance.
After the opening of most positions in August, we will start to submit interviews, and strive to get the offer before April next year.
► Resource Recommendation and Preparation Method:
Here are the 4 most important test centers and recommended preparation resources:
1) Coding skills (including SQL and Python) The content of the database on Leetcode, the main brush easy+medium+small part of hard is enough. I would like to recommend a website called Stratascratch, which contains a lot of real interview questions for big companies. For the Python part, I think the knowledge I learned in school is enough, so I spend time brushing up on the questions every day to keep my hand on it, and discuss it with my classmates.
2) There are several different materials recommended for Product. The first is the Towards Data Science website, which contains many people's insights on product problems, and you can go to see how to solve different product problems when you have time, which is a good help for forming product ideas. The second is 40 Product Questions, which can be searched on an acre and three points of land, and try to read it until it is ripe. There are also many Youtube channels that have very good content. All in all, read more relevant books and videos, communicate more with mentors, and do more mocks!
3) A/B testingMany schools do not have courses related to this aspect, so you need to learn on your own. You can go to Udacity to watch videos, or I highly recommend Emma Ding, a YouTube blogger who has an A/B Testing Series summarizing Udacity's courses on her channel, and the combination of the two can give you a very comprehensive understanding.
4) StatisticsThe question bank of general statistics is very small, you only need to search for the real questions + do it by hand, and in the end, you can quickly grasp the change of the question type and quickly solve it.
More related: What exactly is data analytics doing? Positions & Skills RequiredA full range of popular science DS related positions, required skills, company selection, interview preparation, a full set of gifts! Didn't expect that, right? I finished the class while lying in bed!
► Interview Preparation Advice:
1) Mock Interview can not only improve the ability to communicate with different people, but also improve the ability to respond on the spot. Find one or two friends to form a mock group, and everyone spends some time doing mocks with each other every day, which is a great improvement in interview ability.
2) Product Sense must first have a good mindset: there is no right answer to the product question, and any solution can be a good answer as long as it can be reasonable. During the interview, there is no need to drill into sharp corners, but more how to use communication skills to transform the ideas in your head into a paragraph that the other party can understand, which is the most important ability for the interviewer.
3) Daily exerciseWhen using the product, you can put yourself in the role of designing the product, thinking about the target customer, product benefits and possible defects. With this kind of daily exercise, you will have a broader idea and more different ideas when doing questions, which will help you stand out in the interview process.
4) Think of the interviewer as your teammate rather than a candidate who can give a standard perfect answer, in fact, the interviewer prefers teammates who can communicate with them and have an inspiration for each other. Don't be afraid to make mistakes, because the discussion process is much more than giving a perfect answer.