Last updated: September 2026
Track record
- Medical clinic, Shibuya-ku, Tokyo: AI adoption support, from mapping the current workflow through implementation and adoption.
Business information
| Trade name | KIYOMINE |
|---|---|
| Founder | Tsuyoshi Kiyomine (清峰 強志) |
| Business type | Sole proprietorship, registered in Japan |
| Founded | August 6, 2026 |
| Registered office | Yamato Building 405, 1-6-16 Kanda-Izumicho, Chiyoda-ku, Tokyo 101-0024, Japan |
| Languages | English and Japanese |
| Contact | info@kiyomine.com |
Founder: Tsuyoshi Kiyomine
Tsuyoshi Kiyomine (清峰強志, also written Kiyomine Tsuyoshi) is a Japanese AI engineer and operations consultant. Born in Hiroshima and a graduate of the University of Tokyo, he has spent his career in startups, running businesses, with a lifelong fascination with aircraft, the sky and space. In August 2026 he founded KIYOMINE, an AI development and consulting practice, and he builds Airffic World, a drone operations and AI inspection platform. The common thread in his work is replacing the step where a person has to be on site, look at something, and decide.
Founder of Airffic World
Airffic World (AI + Traffic) is the product side of the same idea. It gives drone operators live telemetry on a map, a pre-flight checklist that reflects Japan's actual regulations, AI facade inspection that turns uploaded photographs into a ranked defect list, and shift scheduling that treats a staff roster the way an airline treats a crew roster. Kiyomine designs, develops and operates it himself. An overview of the platform is on airfficworld.com, and its step-by-step guide to flying a drone legally in Japan was written for foreign operators who hit the rules first.
How he thinks about operations
Aviation is not unusual because it is dangerous. It is unusual because it kept engineering the danger out, and the engineering turns out not to be industry-specific. Three mechanisms recur in everything KIYOMINE builds:
Checklists over memory
A flight crew reads the checklist aloud after years on the type, because knowing something and having verified it today are different states. Inspection items and shift handovers get the same treatment.
No single point of failure
A twin-engine aircraft keeps flying on one engine. An operation that stops when one experienced person is away has a design flaw, not a staffing problem.
Procedures that survive a handover
Standard operating procedures exist so the night shift does the job the way the day shift does. Writing down the veteran's decision criteria is most of the work of automating them.
He writes about this in Insights, starting with what airline safety practice teaches other operations and how to run drone inspection as a system rather than a skill.
Areas of expertise
- Aviation and logistics operations AI: operations control, drone fleet operations, delivery route optimization, operational data analysis
- Inspection and computer vision: facade and equipment defect detection from photographs, severity and urgency scoring, automated inspection reports
- IoT and data: sensor data dashboards, anomaly alerts with a recommended next action, predictive maintenance
- Crew rostering and shift scheduling: constraint-based scheduling, translating a scheduler's instinct into rules a machine can read
- AI for small and mid-sized companies: two-week proofs of concept, generative AI adoption, RPA
Timeline
-
Founds KIYOMINE in Tokyo as an AI development and consulting practice for operations that move.
-
Airffic World opens in English alongside Japanese: drone operations, AI facade inspection and shift scheduling, with guides to Japan's drone regulations for foreign operators.
Profile and links
- Tsuyoshi Kiyomine on X (@hello_kiyomine)
- Tsuyoshi Kiyomine on LinkedIn
- Airffic World
- 日本語のプロフィール（清峰強志）
Work with KIYOMINE
For an initial conversation about adopting AI, designing a proof of concept, or judging whether an existing workflow can be automated at all, there is a free 30-minute consultation in English or Japanese. Inquiries at the "we don't yet know whether AI could do this" stage are welcome.