We operate a large Arabic-language news and content platform. We want to build an AI-powered semantic search and content recommendation system in Arabic.
This requires real expertise in Arabic NLP — Arabic morphology is complex, diacritics matter, and MSA vs. dialect handling is non-trivial.
Deliverables:
– Evaluate and fine-tune Arabic embedding models
– Build a semantic search pipeline using Weaviate or Qdrant
– Deliver a working API with benchmarks
Must have genuine experience with Arabic NLP — not just "I've used Hugging Face before."