AI-Powered Doctor and Nutritionist Recommendation System

Problem Statement

• Finding the right doctor for specific diseases and accessing personalized diet plans is often a complex and time-consuming task. Patients usually struggle to interpret their health reports, and generic diet plans available online don't meet individual needs.There was a need for an AI-based intelligent assistant to bridge this gap with tailored recommendations

Solution

• The system was divided into two core modules:
• Doctor Recommendation System:
• User uploads or inputs a health report (such as blood test results, medical summaries, or symptom descriptions).
• The health report is passed to an LLM.
• The LLM analyzes the report to detect diseases or critical symptoms.
• It retrieves relevant doctors from a vector database populated with doctors’ specialization profiles
• A list of specialized doctors best suited for the user's condition
• Optionally, the system explains the reason for the recommendation
• Input: User provides their age, weight, and height
• Process:The LLM uses this data to calculate BMI and assess dietary needs.
• Process:It retrieves nutrition information from a vector database and generates a personalized diet plan
• Output:A custom daily diet plan including meal suggestions, calorie limits, and nutritional balance tips.

Process / Approach

• Data Collection

  Gathered aerial images of military and civilian vehicles.

• Annotation

  Properly labeled datasets for clear distinction

• Model Training

  Trained a YOLOv8 model & o Fine-tuned for high-altitude, low-resolution images

• Conversion

  Optimized the trained model to HEF using model compression and quantization techniques

• Deployment

  Deployed the HEF model onto drone hardware & Tested in various environments to ensure robustness

System Architecture

• Frontend: Built using Next.js for a fast, responsive, and user-friendly experience.
• Backend: RAG system with LLM integrations
• Database:Vector Database for storing doctor and nutritionist data & Fast semantic search for retrieval.
• LLM: Handles text understanding, analysis, retrieval queries, and generation tasks

Technologies Used

• Large Language Models (OpenAI/Custom Models)

• Chroma (Vector Database)

• Next.js (Frontend Framework)

• Fast Api (Backend API Layer)

Challenges Faced

• Properly structuring health report inputs to extract meaningful information

• Mapping diseases accurately with doctor specializations.

• Generating diet plans that are practical, diverse, and scientifically sound

• Optimizing retrieval performance from the vector database for fast responses.