Skip to content
Asmirai

Push society's availability and adaptability to the limit, and build social systems that are continuously rationalized and optimized.

Turn on-site datainto decisions and working systems.

A data & AI engineering team from Kumamoto University, supporting data use in manufacturing, healthcare, and government — from analytics and data platforms to business systems.

Talk to us about data & AISee case studies

What we build from the data you already have

Your data: Equipment sensor values
What we build: Early anomaly detection
Your data: Sales & inventory records
What we build: Demand forecasts & dashboards
Your data: Free-text survey answers
What we build: Opinion grouping & reports
Your data: Paper / Excel admin work
What we build: Business systems
Typical combinations. We also take analysis-only or platform-only projects.

Case Studies

The challenge, our scope, and what we delivered — shared to the extent we can disclose.

View all case studies
Illustration
Raw dataPoint of interestUsed for inspection decisions

Manufacturing | Sumitomo Heavy Industries Process Equipment Co., Ltd.

Decision Support for Manufacturing Processes through Time-Series Data Analysis

Challenge
Time-series data accumulated on site was underutilized; the client wanted to surface the patterns and trends hidden in it for business decision-making.
Scope of Work
End-to-end: from analysis design to model implementation and insight reporting
Deliverables & where they are used
Anomaly detection and state estimation models, plus a report that puts the findings into words — provided to support decisions in the field.
Read the case study
Illustration
Opinion groupIndividual opinion

Government | Yatsushiro City, Kumamoto / Nagasaki Prefecture

AI-Powered Analysis and Visualization of Citizens' Voices (Broadlistening)

Challenge
Surveys and public hearings captured only a limited range of opinions, and manually aggregating and analyzing large volumes of citizen feedback took enormous time and cost.
Scope of Work
From analysis design to report delivery
Deliverables & where they are used
Classified and visualized opinions, and a report in a form usable for policy-making.
Read the case study

Three areas of support

Ask us for one area, or combine several depending on the challenge.

View service details

Data Analytics & Predictive Modeling

Find grounds for decisions in your data

Typical requests

  • We have data but don't know how to use it
  • We want to forecast failures, demand, or workload
What we do
Data exploration, analysis design and hypothesis testing, time-series / anomaly detection / demand forecasting models
Deliverables
Analysis reports, prediction and detection models, visualizations

Data Platforms & BI

Keep data ready to use at any time

Typical requests

  • Our data is scattered across Excel and multiple systems
  • Aggregation and reporting take too long
What we do
Data collection and integration, ETL pipelines, BigQuery and data marts
Deliverables
Data platforms, BI dashboards, automated recurring reports

Technology & Team

Members with research backgrounds in AI and data science work directly on everything from analysis design to implementation.

More about Asmirai
Yutaro Kuchiki

Yutaro Kuchiki

CEO / Engineer

AI & data science research / full-stack development

A full-stack engineer who has combined AI and data science research at Kumamoto University Graduate School with hands-on implementation in the field. Designs business and technology from three perspectives: research, the field, and the region.

Kazuki Tokuishi

Kazuki Tokuishi

CTO / Engineer

Machine learning & FPGA (PhD candidate) / applied data science

Gained experience as a data scientist at a Tokyo startup and is now a PhD candidate specializing in machine learning and FPGA. Holds a range of achievements in research and AI competitions.

  • 5 members (incl. 2 PhD candidates)
  • Kumamoto University Venture (title granted Dec 2025)
Read the announcement

Technical decisions from published cases

From first conversation to deployment

We combine only the stages you need, based on your challenges and the state of your data.

  1. Consultation

    We hear about your challenges, available data, and operations, and define the scope of the consultation.

  2. Assessment & planning

    We check the state of the data and feasibility, then propose an approach and team.

  3. Technical validationif needed

    We test on a small scale to confirm accuracy, impact, and operational issues.

  4. Implementation

    We build models, data platforms, and business screens, and roll them out in the field.

  5. Operation & improvement

    We refine accuracy, features, and workflows based on how the system is used.

News

All news

No fixed requirements yet? Let's work it out together

Start by telling us these three things, as far as you know.

Talk to us about data & AI
Your data
Equipment sensor values, sales and inventory records, free-text survey answers, etc.
The work that is difficult
Reports are compiled by hand, decisions rely on individual experience, etc.
Where you want to be
Spot early signs of anomalies, check key numbers right away, etc.