Hygieia
A comprehensive healthcare platform deploying 10 decoupled microservices. Hygieia bridges clinical consultations, fitness tracking, and machine learning analysis to provide early screenings and personalized medical advice.
The Challenge
Today's healthcare systems face major challenges: massive delays in treatment, lack of timely guidance, and long travel or waiting times. Fragmentation among diagnostics, fitness apps, and appointment services leaves patients overwhelmed, often skipping or delaying crucial primary and preventive care.
The Solution
Hygieia is a unified, patient-centric digital healthcare ecosystem. It brings professional consultations, active fitness monitoring, lab reporting, and advanced AI-powered diagnostics right to the user's fingertips. By combining 10 microservices, patients receive holistic, timely care in a single application.
System Architecture
Our hybrid microservice setup coordinates FastAPI servers with NestJS gateways to build responsive pipelines.
Project Objectives
Core technical deliverables designed and implemented for the Hygieia system.
Visual Disease Detection
Integrating deep convolutional neural networks to identify dermatological anomalies and classifying dental conditions directly from images.
Stateful Chatbot Companion
Formulating LangGraph-driven conversational flows with dynamic databases, fetching patient history context to resolve medical queries.
Consultation & Lab Bookings
Online booking channels allowing patients to schedule clinical specialist appointments and pathology tests.
Fitness Integration
Collecting live calorie counts, hydration data, sleep patterns, and physical steps by integrating Fitbit APIs.
Asynchronous Alerting
Setting up background schedules, BullMQ task runners, and email notifications to deliver medication updates.
Central EHR Storage
Providing patients with a secure, centralized portal to upload, inspect, and maintain medical records and prescriptions.
Product Features
Advanced modules developed across our frontend and AI services.
AI Health Companion
LangGraph-powered stateful assistant integrated with Groq LLM. It has context of patient records, prescriptions, and appointments. Can book slots or recommend doctors directly through conversation.
Acne & Skin Analysis
Uses an EfficientNet-B3 model trained on extensive datasets. Classifies skin issues into blackheads, whiteheads, pustules, cysts, and papules, providing instant digital screening.
Dental Image Analyzer
Deep learning classifier that inspects dental photos to identify healthy teeth, caries (cavities), and impacted teeth, helping patients detect dental issues early.
Interactive Timeline
A responsive public page mapping the historical evolution of healthcare from ancient Greece (the goddess Hygieia) to medieval medicine and AI-powered remote care.
Fitbit Integration
Real-time activity syncing using Fitbit OAuth. Captures step counts, water intake, sleep quality, and active calories to compute daily personalized health metrics.
Semantic CV Matcher
Recruitment screening service using sentence transformers. Automatically indexes applicant CVs and performs FAISS vector search for semantic candidate discovery.
Project Poster
Our official final year project showcase poster detailing design requirements, systems integration, and product features.
Open high-resolution overview covering scope, technologies, and workflow design.
Technology Matrix
Engineered using modern, enterprise-ready frameworks for robust microservice communication.
Next.js
Responsive and sleek React-based client web platform
NestJS
Structured Node.js gateway and microservices layer
FastAPI
High-performance Python APIs for deep learning inference
RabbitMQ
Asynchronous event broker for inter-service communication
Redis
Powers BullMQ background jobs and reminder schedules
MongoDB
Flexible storage for patient chat logs and session state
Supabase
PostgreSQL storage for application records, users, and transactions
Docker
Containerized deployment for reproducible builds and scaling
LangGraph
Orchestrates the stateful, multi-agent patient chatbot
Groq
Ultra-fast Llama-3 inference provider for recommendations and RAG
FAISS
Performs semantic search over CV documents
PyTorch
Runs skin acne & dental anomaly classification models
Platform Showcase
Explore dynamic interfaces, clinical portal screens, and system analytics via our responsive masonry panel.
Showing 12 of 26 screenshots