CV

Showrov Azam

NLP Researcher | Low Resource Language Processing

Department of Computer Science and Engineering, Daffodil International University, Dhaka, Bangladesh

Email: azam15-5843@diu.edu.bd·Phone: +880 1926 886 230·Location: Dhaka, Bangladesh
Download CV (PDF)View CV in Google Drive ↗LinkedIn ↗GitHub ↗Google Scholar ↗ResearchGate ↗ORCID ↗Web of Science ↗

I am a Computer Science and Engineering graduate (Class of 2025) from Daffodil International University. My research focuses on developing machine intelligence for languages and communities that remain underrepresented in NLP research and technological resources.

My research interests include low-resource language processing, dialect-to-standard normalization, and hate speech detection in low-resource languages. I work on NLP systems that preserve linguistic nuances, code-mixing patterns, and everyday usage of regional dialects instead of flattening them into a single standardized form. My undergraduate thesis applied this approach to regional Bangla dialects, and the underlying methods extend to other low-resource, dialect-rich languages facing similar challenges.

Currently, I am seeking MS or PhD opportunities in Computer Science departments at universities in the United States for Fall 2027, where I can further develop my research in low-resource NLP, dialect processing, language evaluation, and socially relevant NLP applications.

Research Interests

  • Low-Resource NLP & Dialectal Speech Processing: Building transformer-based pipelines that combine fine-tuned ASR with sequence-to-sequence text normalization to convert dialectal and regional speech into standardized text. Resolving register and standardization inconsistencies across low-resource, dialect-rich languages.
  • Endangered and Minority Language Preservation: Developing data curation, annotation, and modeling approaches for underrepresented and minority languages that lack large-scale digital resources, with the goal of extending NLP and speech technologies to communities currently excluded from mainstream language tools.

Research Publications

Full-text access available on ResearchGate, arXiv, and publisher DOI links.

BOISHOMMO: Holistic Approach for Bangla Hate Speech

arXiv preprint · April 2025

Multi-label Bangla hate speech dataset and benchmarks with fine-grained abusive categories. Co-authored with Md Abdullah Al Kafi, Sumit Kumar Banshal, Md Sadman Shakib, and Tamanna Alam Tabashom.

LicenseCC BY-NC-SA 4.0

RGView on ResearchGatearXivDOI

Methodology & Empirical Findings

Key figures, annotation reliability metrics, and algorithmic benchmarks from the paper.

Figure 1 of 5Annotation Reliability
Inter Annotator Agreement Value Distribution over Categories
Inter Annotator Agreement Value Distribution over Categories
Click to enlarge

Per-class agreement proportion among annotators across fine-grained abusive categories including Disability, Physical, Class, Race, Gender, Sexual Orientation, Religion, Ethnicity, Political Statement, and Behaviour.

Quick Switch Figure

Question Generation in Bangla Using Large Language Models: Evaluating Transformer Based Models for Low-Resource Languages

16th IEEE ICCCNT · July 2025

Evaluates transformer-based models for Bangla question generation in low-resource settings. Co-authored with Sayma Jannat, Washik Wali Faieaz, Eftakhairul Islam, Md Sazzadur Ahamed, and colleagues.

RGView on ResearchGate

Methodology & Empirical Findings

Key figures, annotation reliability metrics, and algorithmic benchmarks from the paper.

Figure 1 of 5Corpus Linguistics
Word Cloud for Paragraph
Word Cloud for Paragraph
Click to enlarge

Visual representation of high-frequency Bangla tokens and lexical distribution within the source training paragraphs.

Quick Switch Figure

From Classical to Colloquial: Leveraging LLMs for Sadhu-Cholit Register Identification in Bangla

International Conference on Intelligent Data Analysis and Applications (IDAA) · December 2025

LLM-based approach to Sadhu–Cholit register identification for Bangla. Co-authored with Rasel Parvez, Md Anwar Hossain, Md. Sadekur Rahman, and colleagues.

RGView on ResearchGateDOI

BOISHOMMO: Benchmarking Class Imbalance in Bangla Multi-Label Hate Speech Detection

International Conference on Intelligent Data Analysis and Applications (IDAA) · December 2025 · Daffodil International University, Dhaka, Bangladesh

Benchmarking class imbalance in multi-label Bangla hate speech detection within the BOISHOMMO setting. Co-authored with Sifat Khan, Rashed Hossain, MD Abdullah Al Kafi, and colleagues.

RGView on ResearchGateDOI

Peer Review & Academic Service

Invited peer reviewer evaluating scientific soundness and methodology for international NLP journals and IEEE conferences.

ACM TALLIP

Verified

Peer Reviewer · ACM Transactions on Asian and Low-Resource Language Information Processing (TALLIP) ↗

Invited reviewer for manuscript evaluation, methodology soundness, and clarity of contribution in Asian and low-resource language NLP.

Official Peer Review Verification (PDF)Journal Website

Progress in Artificial Intelligence

Verified

Peer Reviewer · Progress in Artificial Intelligence (Springer Nature) ↗

Invited reviewer for manuscript evaluation, methodology soundness, and clarity of contribution in artificial intelligence research.

Official Peer Review Verification (PDF)Journal Website

ICECCME 2026

Verified

Peer Reviewer · International Conference on Electrical, Computer, Communications and Mechatronics Engineering (ICECCME 2026) ↗

Conference peer reviewer for manuscript evaluation in electrical, computer, communications, and mechatronics engineering.

Official Peer Review Verification (PDF)Conference Website

ICECER 2026

Verified

Peer Reviewer · The International Conference on Electrical and Computer Engineering Researches (ICECER-2026) ↗

Invited to review research submissions for the conference’s technical program, covering emerging developments across electrical and computer engineering, artificial intelligence, communications, signal processing, and related fields.

Official Peer Review Verification (PDF)Conference Website

ACDSA 2027

Verified

Peer Reviewer · International Conference on Artificial Intelligence, Computer, Data Sciences and Applications (ACDSA 2027) ↗

Served as a Review Committee Member evaluating research submissions for the technical program across artificial intelligence, computer science, data sciences, and applications.

Official Peer Review Verification (PDF)Conference Website

Education & Academic Timeline

Undergraduate degree at Daffodil International University with CGPA, academic honors, and thesis.

  • 2022–2025

    B.Sc. in Computer Science and Engineering

    Daffodil International University · Dhaka, Bangladesh

    CGPA 3.68/4.00

    Focused on dialectal Bengali automatic speech recognition (ASR) and preservation of linguistic diversity through deep learning; coursework across algorithms, systems, and machine learning.

    Thesis / capstone

    Final Year Thesis: 'A Two-Stage Framework for Dialectal Bengali ASR and Standardization using Whisper and BanglaT5' supervised by Amit Chakraborty Chhoton. This research addresses the digital exclusion of regional speakers by enabling AI to understand unique phonology and nuances rather than flattening them into a single standard.

    Thesis supervisor

    Amit Chakraborty Chhoton

    Assistant Professor, Department of CSE, Daffodil International University

    Faculty profileGoogle Scholar
    Showrov Azam — Convocation at Daffodil International University

  • 2020

    Higher Secondary Certificate (Science)

    Dinajpur Government College · Dinajpur, Bangladesh

    A+ (5.00/5.00)

    Science group; foundation for university-level CS and quantitative work.

  • 2018

    Secondary School Certificate (Science)

    St. Philip's High School and College · Dinajpur, Bangladesh

    A+ (5.00/5.00)

    Secondary studies with emphasis on mathematics and science.

Undergraduate Thesis

Thesis / capstone

A Two-Stage Framework for Dialectal Bengali ASR and Standardization using Whisper and BanglaT5

Final Year Thesis: 'A Two-Stage Framework for Dialectal Bengali ASR and Standardization using Whisper and BanglaT5' supervised by Amit Chakraborty Chhoton. This research addresses the digital exclusion of regional speakers by enabling AI to understand unique phonology and nuances rather than flattening them into a single standard.

Thesis Supervisor

Amit Chakraborty Chhoton

Assistant Professor, Department of CSE, Daffodil International University

Faculty Profile ↗Google Scholar ↗

Full Report Access

Download Thesis PDFOpen on Google Drive ↗
Undergraduate Thesis Presentation at Convocation

Convocation Thesis Milestone

Work Experience

Flowmingo AI

Recent

Talent Acquisition Specialist

Partnered with technical leadership to identify research-minded engineers and operators; structured interviews and candidate experience aligned with a fast-moving AI product team.

Flowmingo AI certified partner

Volunteering & Mentorship

Take Off

Programming contest organization

Coordinated problem sets, judge workflows, and contest logistics to run fair, engaging competitive programming events for students.

Unlock the Algorithm

Contest operations & outreach

Supported outreach, participant onboarding, and day-of operations to grow the local algorithms community and mentorship pipeline.

Licenses & Certifications

Verified professional credentials in Machine Learning, Deep Learning, and NLP.

University of Michigan · Coursera

Programming for Everybody (Getting Started with Python)

CourseraVerify credential

DeepLearning.AI & Stanford · Coursera

Supervised Machine Learning: Regression and Classification

CourseraVerify credential

Contact

Get in touch

Connect on academic profiles, message on WhatsApp, or write to my university email.