Masayoshi Uehara上原 正吉　—　Engineer & Founder

My first computer was an NEC PC-6601 my parents brought home. Back then there wasn't even a notion of "booting an OS" — you flipped the power and BASIC came up, the screen waiting for input. If you wanted the machine to do anything, you had to write the code yourself. Line by line from line 10, I'd type commands and watch the screen do exactly what I'd told it. That feeling was the start of everything.

Next came Fujitsu's FM TOWNS, one of the earliest machines with a CD-ROM. Before long I was choosing parts one by one and assembling PCs myself. I learned by hand what each part does and where the bottlenecks form — and knowing hardware from the inside became the foundation for designing servers and networks later on.

I went to the US and majored in IT management at university. Not just the technology itself, but how to assemble it into a business and keep it running. Setting both the power to build and the power to make it work as a business as my own axis, early on, is what led to founding a company later.

My first company was founded in America. I assembled PCs that could run both the Japanese and English environments, one at a time, and handled everything from the sale to user support after delivery. Here I learned to see the OS, the network, the server, and what the user is actually struggling with all as one continuous problem. That is the starting point of EarthLink Network Co., Ltd.

From supporting individual machines, I moved to providing the platform that supports a service itself. First I launched a Linux hosting business, then took on Windows hosting while it was still rare in Japan — which Microsoft featured as a case study. You can't take it down, you can't let it slow; in that tension I learned availability and operations in my bones.

I offered rental servers running ColdFusion — then an Allaire product, before either Adobe or Macromedia — among the early ones in Japan to do so, and expanded into dedicated servers as well. Designing the best configuration for each use case became the groundwork for the contract and consulting work that followed.

Requests to "build the system itself" grew, and I moved into contract system development. From there it widened into consulting — which technology to choose, where the risks are, how to build so it lasts — supporting the decisions made before anything is built. From "the person who builds" to "the person who decides how to build," my role changed a notch.

I started at C# 2.0 and worked through ASP.NET and C# 3.0. At the same time I ran a column on C# for All About, putting into words the very places beginners stumble. Building and teaching in parallel — that style took shape around this time.

I held Microsoft and Cisco partnerships continuously. The Cisco Silver Partnership in particular — there were only about 19 in all of Japan at the time, and among big names like NEC and NTT, ours was the only limited company on the list. Being judged not by size but by skill and track record — that feeling still supports my staying hands-on with technology to this day.

As the iPhone began to take off, I saw mobile becoming the next center of gravity and shifted the axis of my work all at once. Before Swift existed, I wrote apps in Objective-C, and replaced every Windows PC in the office with a Mac. That way of "once you're sure, commit everything" maps directly onto how I face AI now.

I moved into Android too and ran a development school teaching both. As gaming was rising and systems written in PHP were moving to smartphones, I moved to Unity, which lets you build once for both. Having worked with it since the 2.0 days, I got calls from company after company, getting down into the floor from adoption support through CI/CD automation.

Before Kubernetes spread, I'd already built a Docker-centered setup. Push a feature branch and an environment dedicated to it stands up automatically, verifiable at its own subdomain like branch.example.com — so each person has their own runtime, free from anyone else's interference. I brought this to companies back when preview environments were not yet common.

At the height of jQuery, I read that React was the next wave and moved my footing there over ten years ago. Beyond implementation, I set coding standards so a team could write to one consistent quality, and designed the onboarding for new engineers. Handing part of the judgment to a mechanism is the very thinking that later carried into drawing the line for what to delegate to AI.

What I've been especially good at is making large systems efficient. Beyond systems with tens of millions of members, there's the spiky traffic common in gaming — leaping all at once at lunchtime, dropping when it's over, leaping again after the workday ends — where designing to the average will always break. I've designed the network architecture that meets the peak yet shrinks without waste.

In every era, I never left what I make to someone else. From the founding I've designed my own company's sites down to the details, and the products and websites I publish today are mine from the look to the internals. I've kept building in-house products too: QuickRec (unified call recording), Heartmap (gaze visualization), WinManager (management and billing for rental-server operators).