Applied AI · B.Eng. thesis · 2024/25
HealthPlatform: a health summary from a medical wearable
An intelligent system that supports health monitoring: it joins biometric data from a medical IoT wristband with the user's own medical documents, and uses AI to generate personal health recommendations.
Goal
Turn data from a medical wristband into readable, personal health insights.
What I built
A React/TypeScript platform with a RAG assistant over the user's own medical documents.
Highlight
Voice in and out: speech recognition for symptoms, spoken answers back.
React 18TypeScriptSupabaseLangflowAstraDBOpenAIGroqElevenLabs
SiDLY bandbiometric data from the medical IoT wristband
SupabasePostgreSQL, auth and storage for medical documents
RAG in Langflowdocuments embedded in AstraDB, retrieved per question
LLMsGPT-4 and Groq for answers and recommendations
VoiceWhisper to listen, ElevenLabs to speak back
Objectives
- IoT device integration: connecting the SiDLY medical wristband.
- AI assistant with RAG: a chatbot that draws on the user's medical documents.
- Recommendation system: machine-learning algorithms that generate personal health advice.
- Voice analysis: speech recognition and symptom analysis from audio recordings.
- Medical documentation: storing and analysing the user's medical documents.
Stack
| Layer | Tools |
|---|---|
| Front-end | React 18, TypeScript, Vite, Tailwind CSS, shadcn/ui |
| Back-end and services | Supabase (database, auth, storage), Langflow for AI workflows, OpenAI API (GPT-4, Whisper), Groq API, ElevenLabs API |
| Data | AstraDB vector database, PostgreSQL via Supabase, Supabase Storage for documents |
Engineering thesis in Computer Science, Warsaw School of Computer Science, supervised by Andrzej Ptasznik, M.Sc.Eng. The code is private; happy to show it in an interview.