LLM-Lab

Ehsaneddin Asgari’s research focuses on developing multimodal and multilingual language technologies, with particular emphasis on document understanding and multimodal retrieval-augmented generation for Arabic and languages of the MENA region. His work is conducted within the Arabic Language Technologies group at the Qatar Computing Research Institute (QCRI).

Research Focus

Our work is organized around two pillars:

1. Multimodal Document Understanding

• Optical character recognition (OCR) for diverse scripts and document types
• Document layout analysis and reasoning over text, tables, and figures
• Retrieval-augmented generation (RAG) — retrieval, reasoning, and generation components
• Domain-specific RAG systems (e.g., Quranic studies, legal texts, historical archives)

2. Multimodal Multidialectal Arabic Language Technologies

• Arabic natural language processing (NLP) across dialects and registers
• Language resources for the dialects and cultures of the MENA region
• Language technologies for digital humanities
• NLP for digital health and well-being

Other Areas of Interest

• Capability-oriented LLM benchmarking and enhancement
• AI for multimedia art and creative applications
• Protein language modeling and antimicrobial resistance prediction using multimodal language models
• NLP for MENA region languages beyond Arabic

Collaborations & Opportunities

We collaborate with a global network of researchers and welcome undergraduate and graduate students in Computer Science and Linguistics for research internships (remote or on-site).

Interested? Send your CV to discuss potential opportunities.