About Me

I am Fariha Anjum Shifa, a graduate student in the Department of Computer Science and Engineering at the University of Dhaka, Bangladesh. I completed my undergraduate degree in Computer Science and Engineering at the same institution.

During my undergraduate thesis, I worked under the supervision of Dr. Upama Kabir on research related to infectious diseases in the field of epidemiology.

My passion lies in advancing natural language processing for low-resource languages, with a particular focus on how large language models understand and reason in Bangla. I am driven to developing methods for training language models efficiently and affordably, no matter which language they serve.

Research Interests

Research Problems I am interested in

• [Low-Resource Understanding]
  How can LLMs truly understand and reason in Bangla and other low-resource languages?
• [Cultural & Social Knowledge]
  How can we model and evaluate culture-specific and socially grounded knowledge in NLP systems?
• [Few-Shot NER]
  How can similarity-based example selection make Bangla NER more accurate and stable with few shots?
• [Efficient Adaptation]
  How can parameter-efficient and multilingual tuning close the gap between high- and low-resource languages?
• [Reasoning Evaluation]
  How do we measure and improve LLM reasoning in cross-lingual, low-resource settings?