Skip to content
View AliMAfana's full-sized avatar
0 followers · 1 following

Block or report AliMAfana

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
AliMAfana/README.md

Hey, I'm Ali

Developer and cybersecurity student building from Gaza. I ship real products, document the journey, and share everything I learn — the wins, the 2 AM bugs, and the architecture decisions that actually matter.

What I'm building

Provia — AI-powered e-commerce platform where every store gets its own AI salesperson. Built with Next.js, TypeScript, Supabase, and OpenAI.

What it does:

  • Upload product photos, AI extracts everything (name, category, selling points)
  • Customers chat with an AI that knows the store's products, policies, and personality
  • Multilingual — customer writes in Arabic, AI responds in Arabic, search stays in English
  • Semantic search with pgvector, session memory, customer profiling
  • Multi-store with separate AI personas per store

Recent writing

  • My AI Kept Recommending Pajamas for Date Night — How vector similarity search doesn't understand shopping intent, and the fallback chain that fixed it. 1000+ followers from this one.
  • The Stock Lie — When the AI agreed with a customer's wrong assumption about inventory
  • 15 Bugs in One Session — Testing marathon that found prompt injection, ALL CAPS crashes, and hallucinated store policies

Tech I work with

Frontend    Next.js 16 · React 19 · TypeScript · Tailwind CSS 4
Backend     Node.js · Supabase (PostgreSQL + RLS) · OpenAI API
AI/ML       GPT-4o · text-embedding-3-small · pgvector · Semantic Search
Security    Rate limiting · CSP · Input validation · Prompt injection defense
Tools       Git · VS Code · Claude Code

The numbers

  • ~550 tokens per chat message (optimized from 1800)
  • 10/10 automated test scenarios passing
  • $0.003 per conversation average cost
  • Built the entire platform in 8 sessions

About me

I'm a cybersecurity student who discovered that building products teaches you more about security than any textbook. Every API I wrote unprotected, every prompt injection I caught, every RLS policy I messed up — those are the lessons that stick.

Building from Gaza isn't a limitation. It's a perspective. Every API call counts when you don't know when the power goes out. Every token matters when your budget is your own pocket. That constraint makes you build better software.

Twitter · LinkedIn · Dev.to

Gaza, Palestine

Popular repositories Loading

  1. AliMAfana AliMAfana Public

    1