fbpx
eWeek Speaks: Featuring Aaron Cheng

eWeek Speaks: Featuring Aaron Cheng

Media Coverage

Aaron Cheng, our Vice President of Data Science, recently spoke with Chris Preimesberger of eWEEK “eSPEAKS” to discuss everything dotData. Watch the full interview here: https://bit.ly/2PVrXwH  #MachineLearning #DataScience

Here is the transcript of the video interview:

Video Transcription

Hi everybody. I’m Chris Preimesberger editor of eWeek. Thank you very much for joining us today on this latest segment of eWeek speaks. So our series of conversations with IT thought leaders from throughout the business. Our interviewee today is Aaron Chang. Aaron is vice-president in charge of data science at dotData.

Chris. Yeah. Aaron, tell me a little bit about your background, just a minute or two, about your background, and then at a high level, what dotData does?

So actually I’ve got a very interesting background. My Ph.D. is actually in physics. I have a Ph.D. degree in physics from Northwestern University, like 14 years ago. So I did physics research for a couple of years and then transitioned into [working with] data. I think data science is a very powerful technology that a lot of companies are embracing right now. That’s how I got into this business. So right now I’m the Vice President of Data Science for dotData, dotData is a software company, developing software solutions for our customers, which enables them to do data science a lot faster to do data science in a lot more efficient way.

Customers who use dotData instead of spending two or three months to work on a product, by using our automation solution, can expect results within two to three days, but that’s what we are doing.

How Companies Use dotData

That’s a good succinct explanation. Aaron, thank you for that. I think it’s interesting that you came from the world of physics into the world of it data, which is science, but it’s a very different kind of science and it’s all about truth.

And numbers and code and data and what to do with it. And so you have joined a company that’s putting that all to work for companies right now. [How are], your customers using, dotData? For what kind of problems are they solving?

Yeah, so basically we provide the analytic software platform. We are not particularly verticalized, towards a specific industry. So from that perspective, I’ve seen some of our manufacturing customers who use our software to develop models, to predict which manufactured parts will fail in the field. I’ve also seen some of the large financial Institute customers [use] our software to determine who [will be] their customer for a [new] product. So there are all kinds of use cases. Again, if it’s a data science-related use case, our tool can be of help.

Very good. So you’re really looking at the past to predict the future in a lot of ways, precisely based interests based on the numbers and the data that you have from the past.

That’s it, [sounds] simple, but it’s not simple [technology at work].

Obviously, that’s why, you know, in today’s market, we all understand data scientists. They are a very, very scarce resource. It’s very hard to find oil. It’s very expensive to hire somebody because you know, doing it right, they can really produce a lot of hours, but the keys to doing it, the right, which is not very easy.

So did you go into data science to make money because they do get paid?

I wouldn’t say that, completely, but it’s one of the reasons, yes, it’s better than being (I don’t want to offend the physicists here.)

Oh, absolutely. Can you name a customer two or three that are using dotData right now?

Yeah, so we have customers, [] Like SMBC. It’s one of the top 15 banks with $1.8 trillion [of] assets in total. It’s, [one of] the largest banks in the world. We also have Toyota [and] manufacturing customers, like Vallureq. They produce a lot of tubing solutions for large oil and gas companies. So they are really from a variety of different industries.

We also have [some] hedge fund customers [who] will use our software to try to address if tomorrow’s market is going to go up and go down.

Wow. Wow. That’s impressive. Toyota. Very impressive. Uh, and all those companies you mentioned, that’s great. How long has that data been in business?

So, dotData, actually [was] part of NEC corporation It’s one of probably the largest IT consulting companies. We started this technology development when we were under NEC I think that’s the right way to put it. We started this whole development, around six years ago [and] became an independent company in 2018.

So from that perspective, as an independent company, we’ve been around for three years.

NEC is one of the largest IT corporations in Japan. Yeah, correct? (Absolutely.) So from your position, you’re, are you the director of data science? Is that pretty much what you do as a vice-president? Are you in charge of products and their functionality?

Most products and also ensuring [that] a customer can get the value and [gathering] the customer’s feedback [for] the products. I take it back to the team, you know, make them part of the product.

Got it. So you really interact with your clients an awful lot and to be sure they’re satisfied with [dotData.]

Correct as well.

Okay. So this, this means that you have to go out on, well, I guess we don’t travel as much as we used to, but you probably do a lot of meetings on video with customers, trying to figure out what their issues are so that dotData can help solve them? Can you give me an example of a problem that you’ve solved recently for a customer that you might want to share with us?

Yeah recently, for example, we’ve worked with a very large retail company. [They] want to do some predictive analytics to try to determine what kind of customers [are] likely to become their next customers for a particular product. It’s a very common type of use case. Understand your customer, offer the right product. So, this is a type of problem we’ve seen a lot because as long as you deal with the customers, you are interested in this kind of problem.

Very good. It’s all about the customer experience now. I mean, with my team from the very beginning is supposed to help us get things done, to buy things, to get services, whatever we need. I think what we’re seeing is we’re seeing more and more granularity in the features and the functions that we’re getting from it.

And yours is very specific and you really help companies solve these issues with customer experience.

That’s what they wanted to do is to provide an excellent customer experience because there’s just so much competition out there.

We have just a couple of minutes left Aaron, but can you think of a trend that you’re seeing in your field, maybe a product or functionality that your customers are asking you for, or some other kind of trend in your field that you’re dealing with?

Yeah, I think if I were to look at what’s happening in the past couple of years, I think that there’s a lot of effort going into just the developing models, going into developing the best machine learning algorithms.

But the new trend I’m seeing is that people are getting more and more interested in deploying these models into production so that they can really utilize [them] to deliver business impact. And this is very different from developing the best algorithm, right? So I’ve seen a lot of customers, they did a tremendous job in terms of putting together a great model that predicts perfectly for the data that they collected. But the challenge is how can they integrate this model into their current IT system so that they can deploy it so that their business folks, every day when they come into the office when they turn on their computer, can see that prediction score shown up on their screen. That’s very different from how a data scientist is running the model and making the predictions.

So that is part of operationalizing data science. Getting the business users to really look at the score, look at the prediction score who react to the prediction score. That is a new trend that I’m seeing across a lot of our customer base.

Aaron, thanks for that. That’s kind of newsy. So what you’re saying is that you’re now developing [is] templates or models?

Yes, we are developing models, but we are also developing tools. To enable them to deliver, to deploy the model, to enable you to integrate the model with [your] IT pipeline.

Yeah, I was just going to say the integration of this is half the battle. Isn’t it? I mean, data integration has always been a major thorn in the side for, many systems, but you guys are figuring out the tools to use, to do this smoothly and completely, and efficiently. Does that include automation?

That includes automation. Absolutely. Yeah. For a lot of manual work that you would otherwise have to [integrate], you know, we are developing a lot of automation tools for that to happen.

Wow, how much time is saved and all that. It’s astounding. We’ll never know exactly cause it’s so much, generally speaking, great stuff.

Aaron Chang, thank you so much for the overview of dotData and, uh, and what you’re doing there. Thank you dotData, if people want to go deeper, is it dotData.com? Right, dotData.com.

Okay. And how long has the company been in business?

We’ve been in business for about three years in the US and globally for around five years.

Okay. Born out of NEC originally. Correct.

Aaron, thank you very much for being on eWeek speaks. I really appreciate it. And good luck going forward. Thank you for everybody following along to the end here. Thank you very much and have a great rest of your thanks for joining us on the eWeek speaks.

dotData
dotData

dotData Automated Feature Engineering powers our full-cycle data science automation platform to help enterprise organizations accelerate ML and AI projects and deliver more business value by automating the hardest part of the data science and AI process - feature engineering and operationalization. Learn more at dotdata.com, and join us on Twitter and LinkedIn.