I am a PhD student in Computer Science at George Mason University, specializing in Machine Learning, Deep Learning, Natural Language Processing, LLMs, Reinforcement Learning, Multimodal Reasoning, and Generative AI.

My research focuses on building reliable, evidence grounded LLM systems especially Retrieval-Augmented Generation (RAG) and agentic AI pipelines that can plan, retrieve, verify, and write with transparency. I specialize in LLMs (RAG, Agentic AI), reinforcement learning, multimodal reasoning, and generative AI , and I work with Dr. Özlem Uzuner on methods that reduce hallucinations by improving retrieval quality and handling agreement vs. disagreement across evidence sources. My recent work includes a multi-source, retrieval-based claim verification direction that explicitly models source-level disagreement using LLMs, along with applied research on LLM prompting for health-related social media understanding aiming to make LLM outputs not only fluent, but auditable and trustworthy.

Recently, I built a multimodal RAG + Deep Research system that achieved 1.5×–1.7× faster responses, 35–40% efficiency gains, and 98% reduction in reasoning time, and I developed a multi-agent writing workflow that reduced first-draft time by 95%. Earlier, I worked as a Software Engineer at BJIT (Tokyo), where I built a personalized recommender/advertising system that reduced manual campaign tuning by 70%.

I am optimistic about making a beautiful world with Science and Technology. I am actively looking for research opportunities and collaborations in my areas of interest and would be delighted to connect with fellow researchers and potential collaborators.

"If an elderly but distinguished scientist says that something is possible, he is almost certainly right; but if he says that it is impossible, he is very probably wrong."

Summer Research Award — George Mason University

Recognition for summer research contribution

3rd Position — 3MT Thesis, Engineering Day

University of Chittagong

Honorary Award — Prothom Alo & Teletalk

Awarded for SSC examination performance

Scholarship — Jessore Education Board

Merit scholarship awarded by the board

Conference and journal reviews across multiple venues in Computer Science

Graduate Teaching Assistant responsibilities at George Mason University

Undergraduate research mentoring and guidance in Computer Science projects

Student volunteer at various international conferences in Computer Science and AI

Outreach activities promoting Computer Science education and technology literacy

Doctor of Philosophy (PhD) in Computer Science

George Mason University, Fairfax, VA

Research Focus: Machine Learning, Deep Learning, Natural Language Processing, LLMs (RAG, Agentic AI), Reinforcement Learning, Multimodal Reasoning, Generative AI

MS in Computer Science — Machine Learning Concentration

George Mason University, Fairfax, VA

Coursework & Focus: Machine Learning, Deep Learning, Statistical Methods, Applied NLP, Reinforcement Learning

Graduate Teaching Assistant

George Mason University — Virginia, United States

Courses: CS 211: Object-Oriented Programming; CS 310: Data Structures

• Led lab sessions and office hours for undergraduate courses; assisted in grading assignments and creating assignments and exam materials.
• Tutored students in Data Structures and Java, improving average assignment scores and helping reduce office-hour wait times.
• Contributed to course material updates and automated marking scripts.

Graduate Research Assistant

George Mason University — Fairfax, Virginia, United States

Worked on key management and network security research; contributed to experiments and codebases.

• Implemented prototype systems for key management and secure communication.
• Assisted in data collection, experimental evaluation, and documentation for publications.

Tech Stack: LangGraph, LangChain, Python, PyTorch, FastAPI, MongoDB, Weaviate, Docker, Kubernetes (GKE)

• Engineered and coordinated a multimodal RAG QA chat system and Deep Research with parallel tool execution for 1.5x–1.7x faster responses and 35–40% efficiency gains; cut reasoning time by 98%.
• Designed and developed a multi-agent writing system with LangGraph and LangChain, improving first-draft writing time by 95%.
• Hardened confidential workflows by deploying privately owned Azure models behind private endpoints and role-based access.

Software Engineer

BJIT Group, Tokyo, Japan

Tech Stack: Spring Boot, JPA, JSP, Amazon EC2, Spiral DB, REST APIs, Python, JIRA, Redmine, JWT

• Engineered a personalized advertising and recommender system for a doctor–patient portal using user signals and session context, improving relevance and reducing manual campaign tuning by 70%.
• Designed and organized the end-to-end data pipeline from Spiral DB to serving layer, developed ETL, computed features, and integrated ML stages feeding REST APIs; deployed on Amazon EC2.

Software Engineer Intern

BJIT Group, Tokyo, Japan

Tech Stack: Spring Boot, Hibernate (JPA), JSP, REST APIs, React, JavaScript, MySQL, Node.js

• Developed a training-management platform for BJIT Academy with real-time messaging and SNS to streamline trainer–trainee communication.

George Mason University
Fairfax, VA, USA