Yu Otsubo
Fullcycle Tech Lead
I'm a Software and Infrastructure developer with over 9+ years, and I've been leading some engineering teams in recent years.
Expert at AWS by using Server-less architecture.
Also a master of Server-side API and Batch with Python.
Additionally, experienced data science with Python and AWS.
I have lectured at AWS Summit Tokyo 19 with company booth.
Skills
AWS
- Lambda
- ECS
- DynamoDB
- SQS
- CloudWatch
- Athena
- S3
- Glue
- Terraform
Backend Develop
- FastAPI(Python)
- Nest.js(TypeScript)
- SpringBoot(Java)
- Echo(Go)
Frontend Develop
- React.js
- Next.js
- Gatsby.js
- Astro
- Vue.js
- Nuxt.js
Database
- MySQL
- PostgreSQL
- ClickHouse
- Snowflake
- Presto with Hadoop
Data Science
- Jupyter notebook
- Pandas
- Scikit learn
- R
- Keras
Native App Develop
- Flutter
- Android Java(WearOS)
Work Experience (8)
Developing and operating some location-based services by using on-premise servers.
Also develop integrated monitoring systems such as Zabbix.
Started the company
Integrate monitoring and operation system using by Zabbix and Expect
I had many experiences such as project management, team building, and handling big data. Our team created the analytics operation system with Splunk and Hadoop(MapR) Used for the company's internal log.
Project management
Team building
Use Splunk
Hadoop(MapR)
Developing and Operating Web-based Chatbot services by using AWS and Python, Node.js.
The architecture is applied by AWS management services such as ECS with Docker, DynamoDB, Lambda, and S3.
Also, creating CI/CD pipelines using by Drone and automation tests using by Pytest.
First experience using AWS
First experience using Python and Node.js
Server-less architecture
Scrum development
Developing data analysis infrastructure using AWS because of analysis of the Chat-bot's conversations.
The architecture is followed by Data-lake architecture.
Data-lake Architecture using AWS
Server-less with Lambda Python
Pandas
Scrum development
Senior engineer and AWS tech Lead of developing Web-based and Native App Chat-bot services by using AWS and Python, Node.js.
The number of Team member is over 8-10, experienced only make architect, developing but also team management, solving difficult issues, and negotiating quite a few stakeholders.
My first experience of tech lead
The Architecture introduced at the AWS Summit
Scrum development
Application engineer of developing Web-based event matching platform by using AWS, Nest.js + React with TypeScript.
Engaged in the full cycle of planning, requirement definition, development, QA, and operation as a full cycle engineer.
Formulate and execute missions and projects for the Product Infrastructure team as Development Manager.
Engaged in the development of a pseudo-live function for live delivery of pre-recorded videos as a video delivery platform/API infrastructure for external collaboration.
As an SRE Engineer, improved continuous development processes using CDK, GitHub Actions, and Autify.
Joined as SRE engineer to build and improve CI/CD pipeline/infrastructure (AWS).
Migration of services built on EC2 to ECS/Fargate while building Blue/Green deployment mechanism.
Expansion of automated testing using Autify and development of execution pipeline to improve efficiency of QA operations.
I develop AI agent applications leveraging generative AI technologies.
Currently serving as the development lead for "A-BOSS" an AI agent created specifically for sales organizations.
Projects (3)
- AWS
- Python
- Elasticsearch
A chat-bot in Web page
Hybrid arch with chat-bot and human operators
Use Elasticsearch for FAQ search engine
- AWS
- Node.js
- Python
A chatbot in Native App
Hybrid arch with chatbot and human operators
Featured at AWS Summit Tokyo
- Amazon Bedrock
- LLMOps
- Generative AI
- AI Agent
- Sales Support System
Developed an AI agent that provides department manager-level guidance to sales teams
Implemented Amazon Bedrock with Agentic Workflow architecture
Created a production-grade LLMOps system with continuous learning capabilities
Integrated with existing sales support systems for seamless workflow
Featured in KDDI's official corporate communications
Education (3)
- Social education
- Educational psychology
- Education within enterprise
- Javascript
- C++
- Ruby
- VisualBasic
- Java
- Unix/Linux
- UML/Design Pattern
- Multimedia(Adobe PhotoShop/AfterEffects/Premier)
- Database(MySQL, Oracle)
- R
- Multivariate analysis
- Social statistics
Awards
Stable and flexible system design utilizing AWS and successfully responded to large-scale traffic without problems.
Publications
This book was written by volunteers from KDDI Agile Development Center, Inc. Each of them brought their experiences and knowledge gained while working on different sites and products, and compiled their practices and learnings as a team into a single book.As a result, the content of this book is not limited to a single point of view, but is a mixture of various approaches and ways of thinking. We would be more than happy if any of the chapters could be a hint for the challenges you are facing now or will be tackling in the future.
An introductory guide to MCP (Model Context Protocol), the standard for the AI agent era. The author of the wildly popular slides “Gentle Introduction to MCP” gently explains the fundamentals of this new technology.
Learn everything about AI agents—the basics, how to build them, and how to use them—all in one book!This comprehensive introductory guide thoroughly explains the fundamentals of AI agents, the hot topic of recent years, step by step from the very basics. It covers essential foundational knowledge, how to build AI agents, real-world use cases, and even Large Language Models (LLMs), which are indispensable when discussing AI agents. The authors are three leading experts in AI application development: Minoru Oda (Tech Evangelist, KDDI Agile Development Center), Yu Otsubo (Software Engineer, KDDI Agile Development Center), and Maki Tsukada (Engineer, AI Strategy Project Group, Mitsubishi Electric). Together, they provide a comprehensive guide covering everything from the fundamentals of LLMs and generative AI to cutting-edge use cases and the latest API implementation methods. The book also provides detailed guidance on how to actually develop AI agents. Reading this book allows you to master the fundamentals of Amazon Bedrock, Claude, LangChain, and LangGraph, enabling you to broadly and thoroughly learn the key technologies for AI agent development. Furthermore, the latter part of the book introduces practical techniques usable in real development environments. It covers AI agent development methods using the recently popular “Mastra” (an AI development framework) and explains post-development operational phase techniques (LLMOps) using Langfuse, taking you directly to a practical level.
Languages
Japanese
English
Interests
Photograph
- Take photos in metropolis night view
- Take photos in Nature
Music
- Original music
Sports
- Watching Baseball
- Playing Table-tennis