Google Cloud & Gemini Architecture

Core Healthcare Infrastructure

Groupe Arex IA Inc. (a Montréal AI health technology startup) engineers its platforms for senior accessibility, patient data security, and high reliability. Our applications are built using Google AI Studio, Gemini multimodal models, Google Cloud Speech-to-Text and Text-to-Speech, Cloud Run, Firestore, Cloud Storage, and Vercel.

Google Cloud Architecture Components:

• Gemini 2.5 Multimodal Vision: Enables instant smartphone camera scanning of meal carotenoid density and AREDS2 formula labels for dry age-related macular degeneration.
• Google Cloud Neural2 Speech Synthesis: Delivers human-grade voice output using en-US-Neural2-F tuned to a senior-paced speaking rate of 0.95x and 0.0 pitch, output in high-bitrate MP3 format.
• Google Cloud Speech-to-Text: Real-time hands-free speech recognition designed to understand older adult speech cadences.
• Google Cloud Run & Cloud Storage: Scalable, containerized backend services ensuring sub-second response times and secure asset hosting.
• Google Cloud Firestore: Encrypted, HIPAA-compliant patient routine persistence and longitudinal adherence metrics.
• Vercel Edge Network: Fast global delivery with static pre-rendering and sub-100ms latency.

Clinical Research Deployment

This architecture powers MacuCoach™ (live product: https://macucoach.com), and is currently deployed in preparation for our 12-week clinical feasibility and usability pilot study in collaboration with the American Macular Degeneration Foundation (AMDF).

Inquiries regarding our architecture or enterprise partnerships: mpiraee@arexai.ca | Founder: Dr. Mahmood Piraee, PhD, PharmD.