Back to Projects
AI Application

MeroDaktar

AI-powered Nepali medical assistant for symptom analysis, voice interaction, and doctor booking.

Tech Stack

PythonFastAPIMed-GemmaLLM Fine-tuningASRTTSReactPostgreSQL

Overview

MeroDaktar is my final year Bachelor project — a bilingual AI medical assistant designed for Nepal's healthcare landscape. It combines a fine-tuned Med-Gemma LLM with automatic speech recognition (ASR) and text-to-speech (TTS) to enable both voice and text interaction in Nepali and English. The system analyzes symptoms, generates structured medical reports, and integrates with a doctor booking flow. All health records are stored securely with role-based access.

Key Highlights

  • Fine-tuned Med-Gemma LLM on Nepali medical datasets for domain-specific symptom analysis

  • Integrated bilingual ASR and TTS enabling fully voice-driven interactions in Nepali and English

  • Automated symptom report generation with structured clinical summaries

  • Secure health record management with role-based access control for patients and doctors

  • End-to-end doctor booking flow integrated with the AI diagnostic layer