Kmuto’s Self-Introduction
Basic Information
| Item | Details |
|---|---|
| Name | kmuto |
| Birth | 1999/04/16 |
| URL | https://eycjur.github.io/resume/en/ |
| X | https://x.com/eycjur |
| Qiita | https://qiita.com/eycjur |
| GitHub | https://github.com/eycjur |
| https://www.linkedin.com/in/katsuhiro-muto-00934b1b7/ |
Skills
Data Science / Machine Learning
- Skilled in data preprocessing and analysis using Numpy and Pandas.
- Experienced with scikit-learn, LightGBM, and Pytorch, and has participated in data analysis competitions.
- Practical experience in anomaly detection using images.
- Extensive experience in developing applications using generative AI, with frequent use of LangChain.
- Interested in MLOps, experimenting with Vertex AI Pipeline and Sagemaker Pipeline as a hobby.
- Selected as an outstanding graduate of GCI 2020 Summer, a data science course offered by the University of Tokyo’s Matsuo Laboratory.
Backend
- Experience in API development using Python (FastAPI), from starting a service (0->1) to maintaining it.
- Familiar with Python (Django, Flask), though without practical experience.
- Experience using Golang.
Frontend
- Experience developing SPA using Vue.js.
- Proficient in developing simple frontends using Python (Dash, Streamlit) and excels at creating prototypes quickly.
Database
- Experience with RDB (PostgreSQL) and NoSQL (FireStore, CosmosDB).
- Knowledge of performance tuning for indexes and queries.
Cloud
- Capable of building a full application on GCP and Azure, though not at the level of a specialized engineer.
- Experienced with IaC using Terraform.
Design
- Strong interest in Clean Architecture and Domain-Driven Design, with a focus on maintainable, extensible designs that minimize technical debt in practice.
Other
Interested in a wide range of technologies and studies them as a hobby.
- Participated in AtCoder, a competitive programming contest, and has knowledge of algorithms. (rating: green)
- Participated in ISUCON, a web service performance tuning contest, and has knowledge about performance tuning. (ISUCON14 134th place/out of 784 teams)
- Interested in native app development and has developed apps using React Native.
- Experience in game development using Unity.
- Interested in IoT, with experience developing using Raspberry Pi.
- Has dabbled in various languages such as C, Haskell, Go, AppleScript, Fortran, and assembly languages.
Work Experience
Leading Data Analysis Company (April 2022 - Present)
- Worked as a machine learning engineer, handling projects involving machine learning and generative AI.
- Particularly adept with generative AI, supporting many clients in chatbot development and generative AI implementation.
- Backend development using Python, frontend with TypeScript (Vue.js), and primary use of GCP/Azure for cloud services.
Public Entries
Activities Outside of Work
Open Source Contributions
Mainly engaged in bug fixes.
Pull Request
- docs: updated args_schema to be required when using callback handlers in custom tools.
langchain-ai/langchain/pull/25887 - fix: error when the environment variable value contains =
smurawski/pycomposefile/pull/31 - community: Fix all page numbers were the same for _BaseGoogleVertexAISearchRetriever
langchain-ai/langchain/pull/19175
Review
- Improvement[VectorStore] Optimize VectorStoreRetriever
langchain-ai/langchain/pull/25091
Issue
- Temperature parameter is ignored when temperature=0
langchain-ai/langchain-google/issues/49
Technical Publications
- Summary of tools for creating projects in Python(Ja)
- Official summary of the 1st class Statistical Examination (Mathematical Statistics)(Ja)
- Seriously thinking about the backpropagation formula(Ja)
- Creating a LINE Notify that sends notifications only on rainy days using the Japan Meteorological Agency API(Ja)
- Tried creating Othello with vue.js(Ja)
- [Python] Divine function to display function argument names in vscode
Other Materials
- Proposal for a Matching Platform Between Investors and Borrowers (GCI2020 Summer Final Assignment)(Ja)