Premalatha

Full-Stack Developer

Chennai Institute of Technologyfull_time, internship

Open to roles

CommunicationProblem SolvingLeadership

MemoryVerse AI '26

MemoryVerse AI is an AI-powered Digital Identity System that transforms scattered academic and professional documents into an intelligent, searchable knowledge repository. Instead of simply storing files, it automatically extracts information, categorizes documents, builds relationships between skills and achievements, generates a digital journey timeline, and enables natural-language search with AI. ### Key Features * User authentication (JWT) * Upload PDF, DOCX, and image documents * OCR-based text extraction * AI-powered document categorization * Metadata extraction (skills, technologies, organizations, dates, etc.) * Interactive knowledge graph showing relationships * Digital journey timeline * Dashboard with profile summary and career insights * Hybrid semantic + keyword search * RAG-based AI chatbot with document citations * GitHub portfolio import * Duplicate document detection * Export generated resume, LinkedIn summary, and timeline PDF ### Tech Stack * Frontend: React, TypeScript, Tailwind CSS, Vite * Backend: FastAPI (Python) * Database: SQLite * Vector Database: ChromaDB * AI: Anthropic Claude API, Sentence Transformers (all-MiniLM-L6-v2) * OCR: Tesseract OCR * Background Jobs: Redis + RQ ### Setup Instructions 1. Clone the repository. 2. Install frontend and backend dependencies. 3. Create a `.env` file and configure: * `ANTHROPIC_API_KEY` 4. Start Redis. 5. Start the FastAPI backend. 6. Start the RQ worker. 7. Run the React frontend. 8. Open the application in your browser and upload sample documents to explore the features. ### Notes * Original uploaded files are preserved and can be retrieved at any time. * AI-generated responses include references to the supporting uploaded documents whenever applicable. * The project demonstrates document understanding, semantic search, knowledge mapping, and intelligent retrieval as required for the MemoryVerse AI '26 challenge.

4 media files · github.comView

MedQ Pro: Live Multi-Doctor Queue & Patient Management System

Patients visiting small and mid-sized clinics often experience long, uncertain waiting times with no visibility into queue status. Reception staff manually manage patient flow, leading to delays, confusion, and frequent interruptions. There was no real-time system for tracking multi-doctor queues or predicting wait times, causing inefficiency for both patients and clinic staff in high-traffic OPD environments Process I started by analyzing real clinic workflows and identifying pain points in OPD queue management. I mapped the patient journey from entry to consultation and found major friction in manual token assignment and lack of visibility. I designed user flows for patients, doctors, and reception separately. Initial wireframes focused on simple queue display, but testing showed users still felt uncertain about wait times, so I iterated to include real-time queue updates and AI-based wait-time prediction. I also explored voice notifications, multi-doctor routing, and priority-based sorting. Some early designs with complex dashboards were dropped due to cognitive overload in testing. Results Improved appointment booking flow from 6 → 2 steps, reducing friction in patient onboarding. Achieved 91% task success rate in usability testing across 14 participants. Reduced perceived waiting uncertainty through real-time queue visibility and AI-based wait-time estimation. Clinic staff workload reduced due to automated queue updates and multi-doctor routing system. Positive feedback highlighted improved clarity and reduced confusion in OPD flow. Reflection If I had more time, I would validate the AI wait-time prediction model with real clinic datasets instead of simulated logic. I would also conduct longer field testing in an actual OPD environment to measure behavioral changes over time. Additionally, I would improve accessibility features for elderly users and explore deeper integration with hospital EMR systems for a more complete end-to-end healthcare workflow.

7 media filesView

This is Premalatha’s work on Wooble.

Build a profile that shows what you can do — and share it anywhere.

Build yours