NanoMind
Open-source AI-powered research co-pilot for nanoscience that unifies literature, experimental data, and domain tools into a single reasoning interface—turning scattered papers into actionable scientific insight.
- Ingested and embedded scientific papers into a vector database (Supabase pgvector) to enable grounded, citation-aware RAG
- Built a streaming chat interface (Next.js + Vercel AI SDK) that answers domain-specific questions with markdown, tables, and structured outputs
- Integrated experimental data workflows (Raman/XRD spectra parsing + visualization) with AI-assisted peak identification and interpretation
- Designed a LangChain agent layer with tool use (Materials Project API, HuggingFace models) for real-time materials data retrieval and scientific reasoning
- Implemented an experiment notebook that extracts and logs structured findings (technique, confidence, source) from AI responses for traceable research memory
ProjectWeb AppNext.jsTypeScriptTailwindSupabaseLangChainAnthropicHuggingFace