AstroWise: Serverless AI Engine & Complex Mobile Architecture
Othmane Zaoute
Project Overview:
AstroWise is not just an astrology app; it is a high-performance computational engine. I architected a system that combines real-time NASA ephemeris data with LLM-powered psychological profiling, all running on a serverless infrastructure. The app handles complex state management, tiered caching, and subscription monetization with sub-second latency.
Key Technical Achievements:
Serverless Architecture: Built 8 specialized Supabase Edge Functions to handle heavy orbital calculations off-device.
AI Optimization: Designed a batch-processing engine for LLM calls (OpenAI/BytePlus), reducing API costs by 85% while maintaining high personalization.
Performance: Implemented a 4-layer caching strategy (Memory → Screen → Date → Database), achieving <100ms load times for daily content.
Zero-Trust Security: Enforced strict Row Level Security (RLS) on Postgres to ensure complete user data isolation.
The Result:
A production-grade, highly scalable mobile platform that processes thousands of data points per user session without compromising UI fluidity.
(See the attached "Technical Whitepaper" for the deep architectural breakdown).
CASE_STUDY.md
Skills:
React Native, Supabase, Edge Functions, TypeScript, RevenueCat, AI Integration.
The Visuals:
Like this project
Posted Feb 13, 2026
Architected a serverless AI engine using React Native & Supabase. Processes real-time NASA data for personalized insights with sub-second latency.
