Skip to main content

Meet SCM Data Scientist Young, who follows Coupang’s logistics journey through data

minute read
IMG 3410

Hello, Young. Please tell us about you and your team.

Hello. I am Young, and I work as a Data Scientist in the SCM Data Science team. 

My team deals with data related to logistics, which is said to be at the core of Coupang, and we particularly focus on data related to fulfillment centers. 

Simply put, we deal with data on almost all aspects of the journey that products sold at Coupang undergo to reach their destination. We handle all data relevant to the product’s end-to-end logistics journey: from delivery by the manufacturer or seller to inbound, loading and picking at Coupang’s fulfillment center, and all the way up to customer delivery.  

The team mainly consists of Data Scientists (DS), Data Analysts (DA), Business Analysts (BA), and Data Specialists. The Data Scientist role is geared more towards the tech side, while the Data Specialist position tends to focus more on the business side. My team mostly consists of Koreans, but we also do have some foreign team members based out of Seattle and Seoul, making it a global team. 

  

What do you do in the SCM Data Science team? 

The team’s Data Specialists are mainly responsible for creating and delivering data or dashboards based on the requests of stakeholders in and out of the Service Design Excellence (SDE) team. In their initial days of joining the company, they work as a team with the DS, DA, or BAs. After they’ve been onboarded to a certain degree, they work as independently as possible under the guidance of BAs to handle stakeholder requests on data/dashboards. BAs also mainly work on creating data or dashboards at the request of stakeholders, but their work is more complex. For example, they are responsible for extracting complex data using queries with thousands or tens of thousands of lines or for maintaining/repairing data marts.  

The DAs have similar roles to BAs in that they find insights from data based on their understanding of the logistics processes built by the stakeholders, but what differentiates them from the BAs is that they are also responsible for more engineering-related roles, such as managing Airflow within the team. 

Now for the DS, we go beyond finding insights from data and leverage optimization (numerical optimization), machine learning, and algorithms to solve the problems faced by stakeholders or improve existing systems. 

 

What was your career path after you joined Coupang as Data Scientist? 

I would describe our team as a team where you can experience Coupang's logistics firsthand working as data experts. 

I don't think there are many companies in the world that can handle logistics data in the depth that Coupang does. In reality, there are only a few companies in the world that can build such an intricate and fast-moving logistics network. In that sense, I think Coupang is currently in an unrivaled position. 

If you work in our team, you will be involved in many projects that deal with one of the world’s best logistics systems in terms of scale and speed. Through such experiences, you will be given the opportunity to go beyond accumulating knowledge in the logistics domain, which is known to have high barriers to entry, and grow into a skilled logistics data expert.  

 

What capabilities are needed to work in the SCM Data Science team? 

Basically, excellent SQL skills are required. Since you need to have a good command of handling data to be able to extract data in the direction proposed by the stakeholders, I would say excellent SQL skills are the most required competency. You also need basic Python skills, and previous experience with BI tools (Tableau, Power BI, etc.) would also be helpful. 

Secondly, communication skills are important. In order to accurately catch and deliver what stakeholders want, you need to constantly communicate with stakeholders and understand their needs. 

By job, the BAs or Data Specialists need the ability to extract data needed by the stakeholders by combining data from multiple sources based on queries of up to thousands or tens of thousands of lines. For the DAs, the ability to find insights from data using Python, etc. is essential, while the Data Scientists need to be able to solve complex problems that occur in various areas of the logistics network based on optimization, machine learning, and algorithms. 

 

Since SCM Data Science team is the only tech team in the SDE organization, is there an organizational culture that is different from regular tech teams? 

It’s in our culture to communicate with stakeholders and collaborate with business teams much more closely than the data team of other general tech teams, and I think this is what differentiates us from the general tech teams. Since we can communicate with business units more frequently when carrying out projects, we can participate more actively in the work being done.  

Also, since our team is a tech team under the SDE team, the work we do often progresses at a much faster pace. Therefore, we work in a more tense environment than other tech teams, which is in line with one of our leadership principles, 'Move with Urgency'. Everything moves quickly in SDE. If there is a good idea that has been verified in advance through data or simulation to deliver value, we spare no resources to test it. To keep up with the speed of SDE, we also must move quickly. 

 

What kind of projects were you able to experience while working in this position? 

You can experience various projects related to inbound, stowing, picking, and outbound at the Fulfillment Centers. 

Coupang uses a heuristic picking algorithm developed in-house based on the Vehicle Routing Problem (VRP), and there is a project aimed at improving/supplementing the existing algorithm by analyzing data related to picking. Another project our team is working on is developing an algorithm that recommends packaging materials based on data for efficient and safe packaging when products are shipped in a single package. 

We also have a project intended at improving the usability of the app used by FC workers by analyzing the mistakes frequently made by pickers based on the picking log of the FCs. In addition to the projects mentioned above, you can experience a variety of projects that can only be experienced at Coupang. 

 

Coupang's coding test is famously difficult, how was your experience? 

As I joined Coupang as a data scientist, I took a coding test that required data handling using SQL and Python. Based on my experience working as a developer and having taken many coding tests, I think Coupang's coding test for the development area is quite difficult. However, the aspects evaluated through the coding test differs across jobs and job groups, and the level of difficulty perceived may vary from person to person, so don’t worry too much about the coding test. I think there won't be a big problem if you make sure to review SQL and basic Python-based data manipulation. 

 

What were your previous jobs and how did you come to join Coupang? 

I worked as a developer and data scientist for R&D at Samsung Research, a Samsung Electronics research institute. At Samsung Electronics, my work revolved around AutoML platform development and ML-based Programmatic advertising (DSP). Then I joined Coupang as a data scientist at Coupang’s FTS(Fulfillment and Transportation System) team, developing picking algorithms, and then moved to Moloco, an advertising DSP company, where I worked as a data scientist. 

Then, I re-joined Coupang’s SDE team half a year ago. The biggest reason for my return to Coupang is its data-driven decision-making culture and the high level of freedom given to data scientists to take initiative at work.  

Many companies publicize that they employ data-driven decision making, but I think there are only a few companies in the world that take data as seriously as Coupang. Most companies take a top-down approach when making decisions, and this is also the case for most tech companies. Coupang also often makes decisions in a top-down manner, but there is one difference. If an idea that comes down from the top can be logically refuted based on data, you can push your point through.  

 

This is related to one of our leadership principles, 'Disagree and Commit', which is impossible to be accepted not only in traditional large corporations but also in most startups or tech companies. Since data is the basis for decision-making, the scope of work that can be performed by data-related roles is much wider than in other companies. We can be involved in work closer to the business side or participate in establishing strategies. Product development or feature development can also be led by the members in data-related positions. These points, which I found very attractive, were what made me come back to Coupang.   

 

What were the parts of your job interviews that seemed differentiated from other companies?  

I think Coupang places a lot of emphasis on communication skills and autonomy compared to other companies, and I believe this is why interviewers at Coupang focus on finding out whether candidates are equipped with effective communication skills and the ability to work independently. In addition to screening the basic skill sets required for a specific job position, they also look at ownership and communication skills for the job. In addition, since Coupang is a global company with talents from various countries, chances are high for you to meet foreign interviewers during your job interview. Interpretation support is provided of course, and while I think such a screening process may incite tension but also provide a fresh and new experience for candidates.  

 

What advice would you like to give to candidates applying for SCM Date Science team? 

On the premise that you have the essential technical skills required of the job, I think the next most important element is owning the attitude of being able to work independently. The interviewer in the interview assumes the role of the stakeholders at work, so I think the key to passing the interview is actively communicating with the interviewers and making them understand the project you have carried out.  

Also, if you are in the interview with interpretation support, and you speak too quickly or ramble because your nervous, it can affect the quality of the interpretation output. Try not to be nervous and speak 'slowly and clearly' so that you can convey your answers much more accurately. 

Tags

Related posts