— Political Science → Tech
Graduated from Keio University (Faculty of Law, Political Science) before joining 42 Tokyo. My background is in social analysis and communication, which I now apply to identifying real-world problems and translating them into working software.
- Search algorithms that resist misinformation and tribalism
- Cybersecurity (particularly supply-chain and national-security adjacent risks)
- Retrieval-Augmented Generation.
RAG-against-the-machine A Retrieval-Augmented Generation system built over the vLLM codebase. Implements language-aware chunking (Python vs. Markdown), a lexical retrieval index (BM25/TF-IDF), and answer generation grounded in retrieved sources using Qwen3-0.6B, evaluated with recall@k.
Seibu-40000-Finder A small tool that pulls live train data from the Seibu Railway to tell you when the next well-seated car will arrive at your station — a personal daily-commute problem solved with a public API.