kmuto.

Software Engineer — RIKEN

Kmuto.

From research to implementation and deployment.
Putting trustworthy AI into practice and turning it into value.

Drawing on four domains — machine learning, software engineering, research, and business — I take AI systems all the way from the first idea to deployment in the field.

I have put machine learning and generative AI into practice across industries such as manufacturing, retail, and finance. When bringing AI into production, I draw on expertise in using it safely, including generative AI quality evaluation and AI security. Today I work at RIKEN on AI for Science, accelerating scientific research with AI.

I also co-authored the technical book "Guide from Senior Data Scientists," contribute to open-source projects such as litellm, and build side products to share what I learn.

(01)

What I Bring

Value

Deep AI expertise

I have been putting machine learning into practice since before the generative AI boom. From deep learning to generative AI and agents, I engage with each problem and reach the accuracy it demands with the method that fits best.

End-to-end development

Ideas don't stall at the PoC stage — they reach the field as working systems. Because I can own every stage from algorithm design to implementation and deployment, nothing falls through the gaps between roles.

Safe AI adoption

AI you can confidently run in production. Drawing on hands-on experience in generative AI quality evaluation, AI security, and governance, I design for risk as well as performance.

(02)

Works & Activities

Works

Publications

Guide from Senior Data Scientists - Engineering Skills to Survive in Practice

Guide from Senior Data Scientists
- Engineering Skills to Survive in Practice

A co-authored technical book that systematizes engineering practices in the data science domain. Gijutsu-Hyohron Co., Ltd., August 2025.

Book page

Awards

  • Outstanding Graduate of the University of Tokyo Matsuo Laboratory's "GCI 2020 Summer"
  • First place, GameDay at the Cloud Native Conference
  • Grand Prize at the Development Organization's Small Improvement Hackathon!
  • Special Award at Nano Banana AI Creator Workshop & CM Challenge
  • AtCoder (competitive programming): Algorithm Green, Heuristic Cyan
  • ISUCON14 (performance tuning contest): 134th/784 teams
(03)

Skills & Expertise

Skills

Data Science / Machine Learning

Hands-on experience from data preprocessing and analysis to ML model building and MLOps.

NumpyPandasscikit-learnLightGBMPyTorchVertex AI Pipeline

Generative AI / AI Agents

Developing and fine-tuning LLM applications. Participated in the Matsuo Lab LLM Development Competition 2025, contributing to foundation model training.

LangChainLiteLLMRAGAI AgentsFoundation model training

Backend

Backend implementation centered on API design and development.

PythonFastAPIFlaskGo

Database

Production experience with both RDB and NoSQL, plus performance tuning knowledge.

PostgreSQLFirestoreCosmosDB

Frontend

SPA development and rapid prototyping. Performance tuning using Lighthouse.

Vue.jsSvelteDashStreamlit

Cloud / Infrastructure

Building applications in the cloud and managing infrastructure as code.

GCPAzureAWSTerraform

Security

Knowledge of AI security; also participate in CTF competitions.

Design / Quality

Design focused on separation of concerns, loose coupling, and modularity. Experience in designing and quality management for enterprise AI products with availability, maintainability, and fault tolerance in mind.

(04)

Work Experience

Career

Research Engineer

RIKEN

Working on AI for Science — accelerating scientific research through engineering, developing AI systems at the front line of research.

Software Engineer & Solutions Engineer

Citadel AI Inc.

Built AI governance products while running pre-sales, PoCs, and implementation support. Learned first-hand that technology only becomes value once it is translated into the language of business. Researched generative AI quality and safety in a NEDO project and published the results as the Generative AI Practical Guide and Case Studies.

Machine Learning Engineer

BrainPad Inc.

Delivered numerous machine learning and generative AI PoCs and systems across manufacturing, retail, and finance. As development lead, drove a generative AI platform from kickoff to release — and learned how to build AI that actually gets used.

(05)

Community Activities

Community

Speaking Engagements

Bridging the Gap Between Theory and Practice
Doc Translate - VSCode Extension for Automatic Code Document Translation Using LLMs
Building a Workflow to Improve the Gen AI App through Prompt Optimization
Proposal for a Matching Platform for Investors and Borrowers (GCI2020Summer Final Assignment)

Many others.

Open Source Contributions

Contribute bug fixes and feature improvements across multiple open-source projects, with particularly active contributions to BerriAI/litellm—including LLM provider integrations and observability enhancements.