Meet CTW’s Senior Data Scientist, Jiang Xiaolan

Path to CTW?

This is part of our series on the team working at CTW.inc.

Today, we’ll spend some time with Dr. Jiang Xiaolan Ph.D., one of our top data scientists. He just goes by Xiaolan, but the man has a doctorate in computer science. We can’t help but call him “doctor.”

Tell us about yourself

Hebei province. It's near Beijing. I came to Japan in 2015 so I’ve been here for seven years now. When I was in university in China, my professor knew a professor in Japan. He recommended him to me for my Ph.D. and he quickly became my supervisor.

My doctoral thesis was in learning-based video streaming. I was working on applying machine learning to improve the viewer’s video streaming experience. I’m quite a traditionally trained data scientist. I learned C, then Java, and then finally Python as I got deeper into the AI side of data science.

Currently, I live in Chiba with my wife. We got married in February, actually!

I enjoy tennis quite a bit and I’ve been playing it for almost ten years. In university, I managed to come in second in a tournament and I love playing it and watching it. Just recently, I saw my favorite player, Roger Federer, retire, which was amazing. He’s had an amazing career, a real legend!

I also like reading books. My goal for the year is to read fifty books. For me, they’re a way to learn about different things in life that I wouldn’t be exposed to without reading. I quite like biographies and books that give me better general knowledge about various fields. I’m interested in quite a lot of things like economics, biology, medicine, and I like to read popular non-fiction books in fields outside of my specialty.?

I really like books on social sciences. I think that’s a very important field today. I like to think about how the impact that AI and data science will have on our society. It’s my hope that we’ll soon be using AI to promote how people prioritize the most essential parts of our society. For example in food distribution, everything from the supply chain to food production, to identifying high-need areas, we can use AI to basically eliminate poverty and hunger. There is a lot of food waste in the world. AI can be used in so many ways, in so many areas to help people become more efficient.

It can change the world.

What surprised you about working at CTW.inc?

After finishing my Ph.D., I worked for a year at Rakuten before coming to CTW.inc. You know, CTW felt like a start-up company, but it’s grown very, very quickly. When I started working here, we had only about fifty people. And now it’s already almost two-hundred-fifty a year later.

So the speed of growth was very quick. And so is the work. I’ve already worked on a huge number of projects. As a data scientist, I’ve grown so much faster than I would have at another company. And I’ve learned a lot, both in my field and about leadership. I’ve been at the company for just over a year but I’m already leading my own team.

I was very surprised when I joined at how freely you could work with the other teams. At Rakuten, if I wanted to work on another project, I had to submit a request and wait for months to work on something which would only take me a week or so to finish. At CTW, if I can finish a project in one week, it’ll get done that week.?

Of course, we still need to align our objectives within the company and we do have big quarterly meetings to set OKRs but there’s a lot of freedom and flexibility within that.

What is your most successful project?

Well, I can tell you about two interesting ones.

The first one is a banner generation model we’ve designed. So let’s say you have a big creative team working on drawing banners in Photoshop. Working hard, a good creative might be able to put out ten banners in a day.? With our banner generation tool, we can create about one thousand banners in one minute.

That project was really big. First, we trained a machine learning model to predict whether a banner would perform well. We gathered a lot of data from the performance of previous banners, ones that we’d placed on Google and Yahoo, as the training data. Using our model, we can now predict what elements of a banner are the most likely to be successful.?

Our banner generation model uses PSD files, which are Photoshop files. We use the model to gather thousands of different layers and elements separately and then combine them in ways that get better and better performance attracting our customers.?

Of course, the model can’t create original images. We need creatives who are very skilled at drawing the illustrations that we use, but they can also modify their images and make slight variations on what’s been successful to find what will perform the best according to our model.

For AI programs that use these machine learning principles, the more data you have, the better the performance. For our most popular games, we have a huge amount of data to modify and train our model. With newer games, there’s not enough banners so all you can do is wait until you’ve put out enough samples for that game to train the model.

The other big project was our G123 translation model. It's our own originally developed model. We’ve compared it to Google’s and ours has better performance, which is very satisfying. Our Global Game Ops teams are now using this model to translate our games for new markets along with our game developers. Liming talked about that in his interview. He worked on the front end, developing the UI, interface and stuff, and I was one of the people working on the functionality part.

What kind of challenges have you faced at CTW.inc?

I’ve not really had any project-related problems, to be honest. Projects are technical challenges and I’ve been solving puzzles for a long time. So that’s fine.

For me, leading a team was a real challenge. It took me a while to feel confident about motivating and guiding my team members. I don’t want to just give instructions. I want to show my people what they can learn and what skills they’re developing from a project. I also encourage them to take on the challenge of owning a project themselves so they can grow as well.

I was lucky that our Chief Data Officer, Wei Liao, was a very good leader. I learned from how he spoke to me and guided me. I basically just took what he did and copy-pasted it to my own leadership techniques. Oh, and I also read some books on effective leadership techniques, trying out a few things to see what worked for me.?

It’s been a great learning experience and I’m lucky because this kind of career growth is much faster than it would be at a bigger company. I think it would take me five years somewhere else to be able to learn as much and develop so many different career skills as I have at CTW.inc!

______

For more information about our available positions, have a look below.?

We’re actively looking for talent in the following roles.

Front-end Engineers

https://apply.workable.com/ctw/j/990993C922/

Backend Engineers: Golang + Python

https://apply.workable.com/ctw/j/7C394159C5/

https://apply.workable.com/ctw/j/6001DCA2DF/

Dev Ops Engineers

https://apply.workable.com/ctw/j/DBCF5492CB/

Korean Game Planner

https://apply.workable.com/ctw/j/7CAD17F7D6/

Traditional Chinese Game Planner

https://apply.workable.com/ctw/j/5FF4A7AD50/

English game or anime IP-related digital marketer

https://apply.workable.com/ctw/j/9B7A842085/

Chinese Version