I'm a PhD student at Hokkaido University (Sapporo, Japan), researching blockchain-based environmental compliance monitoring for mining. Before that: B.Tech + M.Tech from IIT Kharagpur and three software internships building AI backends, VR simulations and data dashboards.
I build backend, data and AI systems: RAG and LLM apps, data pipelines, and anomaly detection for sensor data.
🔎 Open to part-time remote or contract work (up to 20 h/week, JST) · LinkedIn
| Project | What it does | Stack |
|---|---|---|
| rag-doc-qa | Ask questions about your documents and get cited answers. Hybrid BM25 + TF-IDF retrieval, an evaluation set with a CI quality gate, optional Claude answers | Python · FastAPI · scikit-learn · Docker |
| books-etl-pipeline | Scrape → validate → idempotent upsert into PostgreSQL, with retries, run logging and a CI job against a real Postgres | Python · SQL · PostgreSQL · Docker Compose · GitHub Actions |
| mine-sensor-monitor | Finds broken sensors and gas hazards in mine air and water data: caught 254 of 260 injected faults at 0.99 precision | Python · pandas · NumPy · time series |
| pid-hazop-analyzer | Detects components in P&ID engineering diagrams and drafts HAZOP safety reports | Computer vision · Svelte · Node.js |
| stratawatch | Live 3D globe of world mining: mines, minerals, owners, trade flows and prices | JavaScript · data visualisation |
Languages: Python, TypeScript/JavaScript, SQL, C++, C# Backend & AI: FastAPI, Flask, Node.js, LangChain, RAG, PyTorch, scikit-learn, vector databases (Milvus, MongoDB Atlas) Data & infra: PostgreSQL, pandas, Docker, AWS, GitHub Actions Also: Unity VR (welding and mine-rescue training), Power BI
- Starting my PhD at Hokkaido University (October 2026)
- Learning Japanese (日本語を勉強中です)