The programme
Work
Every project Baraa Said has built: products in production, research, and his graduation project. Each one has its own page with how it works and what it achieved.
-
Research
Graph RAG: a 7B model approaching 70B quality
Graph RAG pipelines that turn Arabic and English documents into a Neo4j knowledge graph, so a small 7B model answers close to the quality of a 70B model.
- Neo4j
- LangChain
- Gemini
- Groq
- Pinecone
-
Product · enkiz.org
Faheem: an Arabic AI assistant businesses train on their own documents
Faheem is an Arabic AI assistant that businesses train on their own documents. It answers customers in natural Arabic on the web, WhatsApp, Messenger, Instagram and Telegram.
- Neo4j
- Pinecone
- Gemini
- Python
- WhatsApp API
-
Graduation project
RESCUE-Edge: room-level safety sensing that keeps working when the network fails
RESCUE-Edge is a disaster-resilient edge-AI framework that tells responders which rooms have people and hazards in them, and keeps working when the power and the Internet go down.
- ESP32
- LoRa
- Raspberry Pi 5
- Ascon-AEAD128
- Python
-
Product · nabux.org
NabuX: an Arabic-first learning game of knowledge islands
NabuX is an Arabic-first learning game where every subject is an island. Learners complete short missions, and their world grows with what they master.
- React 19
- TypeScript
- Phaser
- Node.js
- SQLite
-
Product · dukkan.fr
Dukkan: a marketplace for independent boutiques
Dukkan is a multi-vendor marketplace where independent boutiques open their own shop and sell online, while the platform handles payments, seller payouts and shipping.
- Next.js 16
- React 19
- Prisma
- PostgreSQL
- Stripe
-
Research
Can classical speech features still catch AI-generated voices?
Research testing whether classical speech features can still catch AI-generated voices. Across 69 detection systems, generalisation to unseen attacks and codecs, not the choice of features, turned out to be the real bottleneck.
- Python
- MFCC
- LFCC
- CQCC
- SVM
Questions?
Hiring for an AI or full-stack role, or need a system built? One message is enough.