Loading
Nader Bennour

Senior AI & LLM Engineer

RAG Systems Builder

AI Infrastructure Architect

  • About
  • Works
  • Services
  • Resume
  • Skills
  • Blog
  • Contact
Nader Bennour

Senior AI & LLM Engineer

RAG Systems Builder

AI Infrastructure Architect

Download CV

Recent Posts

  • RAG in 2026: Why Most Pipelines Still Fail in Production
  • Multilingual LLM Systems in 2026: What Changes When Your AI Needs to Speak 4 Languages
  • From 50 Seconds to 3: Cutting LLM Inference Latency in a Production RAG System
  • The Agentic AI Gold Rush: Why Engineers Who Ship to Production Charge 3x More

Recent Comments

No comments to show.

Archives

  • January 2026
  • November 2025
  • September 2025

Categories

  • AI Career
  • AI Engineering

Vitalyze — Multi-Agent Health Intelligence

  • Created By: Nader Bennour
  • Date: 2025
  • Client: Figurmacher Academy
  • GitHub: View Repository
Vitalyze is an AI-powered precision health platform. Upload lab results, genetic raw data (23andMe/AncestryDNA), or medical documents — the system routes everything through 7 domain-specific AI specialists running in parallel, cross-validates findings across sources, and delivers a comprehensive health assessment with confidence scores.
The pipeline filters 586K+ raw SNPs down to 120+ clinically actionable genetic variants using real-time NCBI dbSNP annotations. Each specialist agent (blood, genetic, hormone, nutrition, training, symptom, methylation/detox) analyzes its domain independently, then a master synthesis agent cross-validates and resolves contradictions. Zero data is stored — everything auto-deletes within 24 hours.

Technical Highlights

  • 7 parallel specialist AI agents with domain-specific prompting
  • Cross-validation engine that scores confidence across data sources
  • Genetic analysis — 586K+ SNPs filtered to 120+ actionable variants via NCBI dbSNP
  • Document processing — PDFs, lab images (German format), spreadsheets via Mistral OCR + EasyOCR
  • Interactive health coaching chat with Eleven Labs voice input
  • Privacy-first — zero persistent storage, 24h auto-deletion

Stack

  • AI: OpenAI GPT-4o, Google Gemini 2.5, Mistral AI (OCR)
  • Backend: FastAPI, Python, Uvicorn
  • Genetic: NCBI dbSNP E-utilities, SQLite caching
  • Document Processing: PyMuPDF, EasyOCR, Pandas, OpenCV
  • Speech: Eleven Labs STT
Tags: FastAPI Gemini Health Tech Multi-Agent AI OpenAI
Share:

Prev
Next

© 2026 Nader Bennour. Senior AI & LLM Engineer — nader.info