Takeuchi: What kind of boy were you when you were young?

Nishikawa: I became interested in PCs around the fourth grade. My father had an introductory book on (the programming language) BASIC. I picked it up and I was immediately hooked. My parents wouldn’t buy me a PC, so I would play around with source code on a piece of paper.

After I got my own PC, I spent days programming. Then, I gradually became curious about the more fundamental aspects of computing, not just programming — like how computers work in the first place? From there, my interest grew in the direction of networks and hardware. I didn’t think twice about joining the Hiraki lab (professor Kei Hiraki’s lab at the University of Tokyo’s Graduate School of Information Science and Technology).

Takeuchi: It’s an interesting lab. I hear the professor still visits your company.

Nishikawa: Yes, he does. Actually, he works for us full time.

Takeuchi: So, you asked him to join PFN?

Nishikawa: No, he first told me he wanted to talk about something and visited our office. Until then, he had advised us on analogue aspects, like the handling of processor boards and high frequencies and things like that.

One day, he told us he wanted to develop a supercomputer with PFN. He now gives lectures to our members once a month or so.

Takeuchi: Am I correct in remembering that you met (PFN co-founder and Vice President Daisuke) Okanohara in college?

Nishikawa: Correct. We were in the same class in the first year, and our relationship has continued to this day.

He said he wanted to run an application on a search engine he developed through the MITOU Program. While I helped him with that, we started discussing our dream of creating a sophisticated search engine and launching a business around it.

Takeuchi: So, that led you to establish (PFN predecessor) Preferred Infrastructure (PFI), right? I understand you launched the business when you were in graduate school. What were your career plans at the time?

Nishikawa: At the time, I was interested in IBM or Intel. While I was working part-time for a bio startup, this experience helped me realize it was possible to work with a small team that enabled me to do and approach things I wanted to in ways I liked. That’s why I decided to start a company, but I had little funding. Starting a company wouldn’t have been possible without MITOU’s support.

Support in early years

Takeuchi: So, the business took off smoothly?

Nishikawa: No, not at all. The founding members were all engineers. We had zero knowledge about making money, let alone how to establish a company.

That’s where MITOU’s support proved helpful. The program referred us to experts who could give us advice about things involved in starting a company, like legal, financial or intellectual property matters. An accounting firm that helped us back in those days is still doing some work for us.

Takeuchi: It’s been 10 years since we last had time for this kind of discussion. Your company has grown so much during this period, hasn’t it?

Nishikawa: Fourteen years ago, Preferred Infrastructure — now PFN — was created by six people, including team members from an international programming competition, as well as my university friends.

We established PFN in 2014 to take advantage of our real-time machine learning technology that focused on the “internet of things.” We are now fortunate enough to have about 300 employees.

Takeuchi: Wow, six of you started the company and it now has 300 employees. That’s impressive. (PFN co-founder and Vice President Daisuke) Okanohara was in MITOU. Do any other former MITOU participants belong to your company?

Nishikawa: We have at least 13 MITOU “graduates.”

Takeuchi: I think that may be the largest number of MITOU graduates now working for a single Japanese company. You have grown the company through identifying outstanding individuals and putting them together as a working team. I think that deserves praise.

I’ve been involved in MITOU as senior project manager. I notice many applicants mention they “did an internship at PFN.” What does it take to successfully join your company?

Nishikawa: First of all, we look for someone who is knowledgeable about computer science. We are divided over whether we should give a coding exam to applicants, but in my opinion, we should. That’s because I think it’s important to find out not only whether a candidate understands theory, but also if they can also practically apply said theory.

For applicants who interned with us over a two-month period, we review how they performed during their internship to find out how they identified issues and how they addressed challenges.