Kurdish NER with XLM-R
01Fine-tuned xlm-roberta-base for Named Entity Recognition in Kurmanji Kurdish. Achieved an F1 score of 0.87 on the test set.
View ModelSpecializing in machine learning and natural language processing.
Learn moreMachine learning engineer focused on NLP and applied AI.
I work across the full ML lifecycle, from data collection to deployment. Currently focused on Kurdish language processing.
Ovanya
Building AI solutions with focus on machine learning and computer vision.
UKH AI Center
Developed NLP models for Kurdish language processing. Collected and processed 140K+ news articles for training data.
Standing Tech
Built and maintained websites for local businesses.
University of Kurdistan Hewlêr (UKH)
Focus on software development, algorithms, and AI.
Ovanya • wedo
Intensive program on ML fundamentals and neural networks.
Fine-tuned xlm-roberta-base for Named Entity Recognition in Kurmanji Kurdish. Achieved an F1 score of 0.87 on the test set.
View Model
A web application that identifies whether Kurdish text is written in Sorani or Kurmanji. Achieved 92% classification accuracy.
Live Demo
A web tool that converts Kurdish text between Arabic and Latin scripts with accurate transliteration.
Visit Website
A platform for discovering and downloading open-licensed Kurdish fonts for content creators.
Visit Website
Collected and processed over 140,000 Kurdish news articles for NLP research at UKH AI Center.
Peer-reviewed paper presented at the 9th Widening NLP Workshop (WiNLP 2025), demonstrating an XLM-RoBERTa-based NER system for Kurmanji Kurdish with strong F1 scores.
Read paperInterested in working together?
akam.nawzad@ukh.edu.krd