Applied Team Member Introduction

Sakana AI has quietly assembled a new Business Development Division and, in typical understated fashion, introduced it to the world not with a splashy manifesto but with a slow-burn series of staff profiles. The Tokyo-based lab, known for research into evolutionary algorithms and self-improving models, says the unit exists to turn its R&D output into actual products for Japanese corporations and, notably, for defense and public-sector clients.

That defense angle is the part that stands out. Chieko Tazuke, who moved from Japan's Ministry of Economy, Trade and Industry through deep-tech investing at Sony Ventures, now splits her time as both a defense-focused project manager and de facto chief of staff for the whole Applied Team. She's blunt about the challenge: most frontier models assume constant cloud access, and that assumption falls apart the moment you try to put AI on a vehicle or piece of equipment operating without a network connection. Her stated interest is in small, function-specific models built for exactly that constraint — a very different problem than scaling up another giant LLM.

Elsewhere on the team, the profiles read like a tour through Japan's tech-to-AI pipeline. One engineer came from Mercari's fraud-detection systems and later built recommendation and demand-forecasting pipelines at a US e-commerce startup before landing on full-stack engineering at Sakana, where AI coding agents now handle a chunk of the grunt work. Another, Haruki Goda, spent his early career at Omron doing image-processing software before pivoting to building AI agents for enterprise clients, chasing what he calls