Home / Resources
RESOURCES
Insights on Arabic AI & Language
Field notes, research, and best practices from our Arabic AI and linguistics team.
Arabic NLP
Why Diglossia Breaks Most Arabic Language Models
A look at how MSA-only training data causes models to fail on real spoken Arabic.
Jul 21, 2026
Model Evaluation
Designing Rubrics for Arabic RLHF at Scale
How we structure preference ranking guidelines across 20+ dialects.
Jul 7, 2026
Dialects
A Field Guide to Gulf vs. Levantine Arabic for AI Teams
Key vocabulary and grammar differences that matter for annotation accuracy.
Jun 23, 2026
Annotation
Building Inter-Annotator Agreement Into Arabic QA
The statistical methods we use to keep large annotator teams consistent.
Jun 9, 2026
Arabic NLP
Right-to-Left Text: An Underrated AI Engineering Challenge
What breaks when tooling assumes left-to-right script.
May 26, 2026
Model Evaluation
Prompt Evaluation Lessons From 1M+ Arabic Responses
Patterns seen reviewing Arabic model outputs at scale.
May 12, 2026