Abdullah Ibne Hanif Arean

I am an AI researcher at the University of Dhaka, where I completed my B.Sc. in Computer Science & Engineering at the top of my class and am now finishing an M.Sc. thesis on streaming graph recommendation. My research focuses on trustworthy AI, multimodal reasoning, graph learning, and low-resource NLP, so intelligent systems are not only powerful, but also explainable, robust, and fair.

In practice, that means asking hard questions of modern models: when do vision-language systems invent structure that is not there? How do we evaluate safety when languages like Bangla, Sinhala, or Yoruba are under-served by English-centric benchmarks? How can recommenders adapt over time without erasing what they already know? My papers and ongoing work, spanning venues such as ICDAR, AACL-IJCNLP, and submissions to WACV, ICLR, and NeurIPS, try to answer those questions with methods that can actually be audited and reused.

My life goal is simple to say and hard to finish: help more people benefit from AI without being harmed by it. I care especially about communities and languages that are usually left behind by large-scale AI progress. That goal shows up in four habits: publishing careful research, building systems that leave the lab, mentoring the next generation of researchers, and strengthening the tech community around me.

Along the way I co-founded ARIBD to grow a self-sustaining, mentorship-driven research culture in Bangladesh, and Sigmoix AI to bring research-backed AI into enterprise use. Through teaching, workshops, and student leadership (including roles such as President of the Dhaka University IT Society), I have reached more than 2,000 learners. I am always open to collaborations that make AI safer, more useful, and more accessible.