[Event Report] Applied Engineer Open House 2025: The Frontline of AI Social Implementation Tackling Difficult Challenges in Finance and Defense
Sakana AI ● Covered by 3 sources
Sakana AI held its first Applied Engineer open house, showing off how it's tackling finance and defense AI projects in Japan. It's a rare peek at how a research-heavy AI lab turns papers into real bank and military tools.
Based on reporting by Sakana AI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Tokyo-based Sakana AI threw open its doors on August 7th for a first-of-its-kind event aimed squarely at engineers, not investors or press. Roughly 70 people showed up in person, another 200-plus tuned in online, and the Q&A session pulled in more than 50 questions — a sign that the company's pitch about doing unglamorous, high-stakes AI work in Japan is landing with the people it wants to hire.
Co-founder and CEO David Ha opened with the origin story: a Canadian who moved to Japan over a decade ago, cut his teeth at Google during deep learning's boom years, and grew frustrated watching AI leadership concentrate almost entirely in the US and China. His argument is straightforward — Japan, as Asia's leading democracy, ought to have a seat at that table, and Sakana AI exists to put it there. The company still publishes research openly, but its actual revenue engine is built around selling AI products to large Japanese corporations and government bodies, a deliberate bet that a small startup creates more impact by going deep on a few hard problems than by spreading thin.
COO Ren Ito, a former diplomat of 15 years before he jumped into startups, made the sharpest point of the day: building a ChatGPT clone doesn't actually fix Japan's problems. Automating a bank's loan approval process, he said, means burrowing into that bank's actual workflows and grinding toward 99% accuracy — engineering that's far less glamorous than publishing papers but arguably more useful. Sakana AI already has major banking clients signed and is now pushing seriously into defense work too, an area the company sees as both underdeveloped in Japan and a genuine technical proving ground.
Several engineers filled in the texture. Ryuichi Kanoh talked about a US-Japan defense innovation competition Sakana AI won, framing defense as a space with genuinely hard technical problems and few AI startups willing to touch it. On the finance side, Hiroyuki Ota described clients demanding
My take — AI-written commentary, not fact-checked reporting
What stands out here isn't the tech, it's the confidence to say the quiet part out loud: another chatbot doesn't move the needle, but a loan-approval pipeline running at 99% accuracy for a megabank actually might. I'd rather see more labs chase boring, high-stakes plumbing like this than another benchmark-topping demo, and Japan going all-in on defense AI while everyone else tiptoes around the topic is the kind of bet that either looks prescient or reckless in five years — probably both.
Read more about this at: Sakana AI
