Md Tanzib Hosain

I am a fresh graduate student with a bachelor's degree from the Department of Computer Science, part of the Faculty of Science and Technology at American International University-Bangladesh, where I worked on computational linguistics integrated human computer interaction and federated machine learning.

I have worked as a research assistant at Qatar Computing Research Institute, as part of Knowledge Augmentation. My research centers on AI for mathematical and scientific understanding, spanning (1) Language Models—LLMs as search agents, multimodal pre-training, generative models for mathematics, LLMs for code generation and scientific discovery, and reinforcing language agents for reasoning—and (2) AI4Science, including transformer and graph neural network approaches to learning algorithms, machine learning, human-computer interaction, and optimization for scientific discovery.

Email  /  GitHub  /  Google Scholar  /  LinkedIn

Intrinsic Linguistic Bias in Formal vs. Informal Bengali Pragmatics with Progressive Context Inflation

We studied observational measures of intrinsic gender bias in formal and informal bengali; the optimal context length influences on bias detection.

B-REASO: A Multi-Level Multi-Faceted Bengali Evaluation Suite for Foundation Models

We introduce a bengali language benchmark for large language models, covering 13,497 multiple-choice questions in 50 different subjects, divided into 4 levels of difficulty.

Can Multi-turn Self-refined Single Agent LMs with Retrieval Solve Hard Coding Problems?

We present a benchmark which consists of world finals’, continentals’ and regionals’ international collegiate programming contest problems.

𝕏olver: Multi-Agent Reasoning with Holistic Experience Learning Just Like an Olympiad Team

We introduce 𝕏olver—a training-free, multi-agent reasoning framework that equips a black-box LLM with a persistent, evolving memory of holistic experience.

Multimodal Programming in Computer Science with Interactive Assistance Powered by Large Language Model

We present initial implementation of an interactive homework assistance system, for students of introductory computer science programming course.

A Hybrid Self Attentive Linearized Phrase Structured Transformer based RNN for Financial Sentence Analysis with Sentence Level Explainability

We Introduce an interpretable, attention based RNN for linearized phrase structured sentence sentiment.