Harshit Bisht

MemoryVerse AI '26

Project Name: Digital Identity — AI Academic & Career Journey Platform GitHub Repository: https://github.com/Harshit17x/Digital-identity Tech Stack: React 18, Vite, Google Gemini 2.5 Flash, Supabase pgvector, Vis.js Graph. Key Technical Highlights NLP & Entity Extraction: Auto-extracts technical skills, issuers, dates, and confidence scores (0.85–0.99) from uploaded PDFs/documents via Gemini 2.5 Multi-Modal & token matching. RAG System & Chatbot: Answers natural queries grounded in user credentials with cited document metadata using Gemini 2.5 Flash. Embeddings & Vector Search: Implements L2-normalized vector frequency embeddings and cosine similarity search inside an instant Cmd + K Command Palette. Vector DB Schema: Includes supabase_schema.sql configuring PostgreSQL pgvector (embedding vector(768)). Knowledge Graph: Vis.js network graph connecting category hubs, documents, and cross-domain skill clusters. Quick Testing Steps for Evaluators Landing Page: Opens directly on the public hero page. Upload Hub: Drag & drop any certificate or resume to test automatic AI parsing. Timeline Page: Use the sidepanel sliders to filter by year or AI confidence threshold. Command Palette (Cmd + K): Search by skill (e.g., PyTorch, AWS) for instant semantic match rationales. Pricing Section: Displays plan features with a notification that payment gateway integration will be added in a future update.