December 2024 - February 2025
This repository documents my work as an AI Developer at AVILEN, where I implemented an automated research system based on the STORM architecture. The system generates comprehensive research reports in Japanese by gathering and synthesizing information from academic papers, patents, and web sources.
- Led evaluation of various AI technologies and search APIs
- Created systematic assessment framework
- Documented capabilities and limitations of different approaches
- Implemented STORM architecture with Japanese language support
- Created separate implementations for different data sources:
- Academic paper research using OpenAlex API
- Patent information search using SerpAPI
- Web information search using DuckDuckGo
- Documented system design and results in both Japanese and English
AVILEN-linkers-project-portfolio/
├── images/
├── notebooks/
│ ├── Topic3_Scholarly_Research_STORM.ipynb
│ └── Combined_Search_Implementation.ipynb
├── README.md
├── technology-evaluation/
│ └── README.md
└── implementation/
├── README.md
└── experimental-outputs/
- Japanese language support throughout the system
- Modular search implementations for different source types
- Multi-perspective research approach
- Automated citation management
- Structured data handling using Pydantic
- AI/ML: LangGraph, OpenAI GPT models
- Search APIs:
- OpenAlex API for academic papers
- SerpAPI for Google Patents
- DuckDuckGo for web search
- Development: Python, Pydantic
- Documentation: Bilingual (Japanese/English)
System tested with topics including:
- エクソソームの研究のトレンド (Exosome Research Trends)
- アミンを用いた固体状のCO2吸着剤のトレンド (CO2 Absorption Technologies)
- 熱・風以外の方法で髪の毛を乾かす技術のトレンド (Novel Hair Drying Technologies)
- 安全に関する行動変容手法のトレンド (Behavioral Change Methods)
- 加熱調理器における料理品質・体験・機能に関する開発のトレンド (Cooking Appliance Innovations)