S M Jishanul Islam Webpage

About

Research that leaves the lab.

I am a master’s student in Computer Science at Dalhousie University, supervised by Prof. Dr. Frank Rudzicz, and an AI Engineer at pipikwan pêhtâkwan, where I train and serve AI built to corrects unconscious bias or racism directed to Indigenous Peoples in written materials.

Before Canada I did my undergraduate in Computer Science and Engineering at United International University in Dhaka, majoring in Data Science - a perfect 4.00 in the major, Cum Laude overall — and spent a year building production OCR and vision-language systems at Apurba Technologies. I used to also teach at my Alma Mater, part-time, because the questions students ask are the ones worth answering.

My published work runs from tri-modal adversarial attacks on short video to Bangla OCR on documents nobody else would touch. The through-line is the same: several signals, one understanding. During my master's, I am trying to make medical AI understand hierarchy in medical ontologies and help them ACTUALLY see, read, and hear.

Education

Where I learned it.

Master’s in Computer Science
2026 — Present

Dalhousie University Halifax, Nova Scotia, Canada

Supervisor: Prof. Dr. Frank Rudzicz
B.Sc. in Computer Science & Engineering
2020 — 2024

United International University Dhaka, Bangladesh

Major in Data Science · Cum Laude · Supervisor: Prof. Dr. Swakkhar Shatabda

3.77/4.00CGPA

4.00/4.00Major CGPA

Experience

Where I’ve built things.

AI Engineer
Oct 2025 — Present

pipikwan pêhtâkwan Edmonton, Canada (remote)
- Trained LLMs for multi-purpose tasks serving Indigenous Peoples, reaching a state-of-the-art 86% BERT F1.
- Built inference pipelines for billion-parameter models at scale, cutting time-to-first-token to 1 second.
- Led the training of the first thinking model built to serve Indigenous Peoples.
- Work alongside the software engineering team to fold LLM pipelines into AI-powered sub-products.
Part-time Faculty
Oct 2025 — July 2026

United International University Dhaka, Bangladesh
- Teach core applied CS courses as faculty from industry, centred on algorithms and software engineering.
- Work with the CSE department on grading, feedback and performance reporting, bringing industry perspective back into the courses.
- Coordinate with TAs on course materials and outline design.
Software Engineer
Dec 2024 — Oct 2025

Apurba Technologies Ltd. Dhaka, Bangladesh
- Built scalable, optimised core ML services, improving performance by 10%.
- Co-created the first CLIP-style word recogniser for VLM-based Bengali OCR, at 95%+ accuracy.
- Worked with stakeholders to research and deliver project milestones two to three weeks ahead of deadline.
Lecturer (contractual)
Jun 2024 — Oct 2024

United International University Dhaka, Bangladesh
- Taught core CS concepts, raising average student performance by 7% through examination and targeted feedback.
- Courses: Object-Oriented Programming Lab, Data Structures & Algorithms I Lab, Database Management Systems (Theory).
Undergraduate Assistant
Sep 2023 — May 2024

United International University Dhaka, Bangladesh
- Facilitated course content with lab faculty, evaluated assessments and judged lab projects.
- Wrote the first course materials for the country’s first undergraduate degree in Data Science.
- Courses supervised: Programming for Data Science, Object-Oriented Programming for Data Science.

Publications

Answering one question.

How do you get a model to understand something when the evidence arrives in more than one form?

Journals

Q1 IF 8.0

An audio video-based multi-modal fusion approach for speech emotion recognition

Islam, S. M., Mustakim, S. H., Hossain, M., Mashira, M., Shourav, N. I., Ahmed, M. R., Islam, S., Islam, A. K. M., & Shatabda, S.

Knowledge-Based Systems (Elsevier) · 2026

Conferences

CORE A

BornoDrishti: Leveraging Vision Encoders and Domain-Adaptive Learning for Bangla OCR on Diverse Documents

Islam, S. M., Hassan, M., Ovi, M. H., Rabby, A. K. M., & Rahman, F.

19th Conference of the European Chapter of the ACL (EACL), Industry Track · ACL, 2026
CORE B

Beyond Characters: Position-Aware Metric and a Lightweight LLM for Low-Resource Bangla Punctuation Restoration

Hassan, M., Islam, S. M., Rabby, A. K. M., & Rahman, F.

IEEE International Conference on Big Data (BigData) · IEEE, 2025
CORE C

IPBlocks: A Blockchain Ecosystem for Secure IP Registration and Decentralized Marketplace

Ahmmed, S., Islam, S. M., Mustakim, S. H., Islam, R., Shanto, S. K., & Islam, S.

TENCON 2025 — IEEE Region 10 Conference · IEEE, 2025
E-MedViTR: Enhanced Vision Transformers with Registers for Biomedical Image Classification

Ahmmed, S., Rahman, T., Islam, S. M., Reyad, A. M., Dey, S., & Farid, D. M.

6th International Conference on Electrical Engineering and Information & Communication Technology (ICEEICT) · IEEE, 2024

Workshops

CORE A* workshop

Watch, Listen, Understand, Mislead: Tri-modal Adversarial Attacks on Short Videos for Content Appropriateness Evaluation

Islam, S. M., Mustakim, S. H., Ahmmed, S., Sayeedi, M. F. A., Khandoker, S., Dhrubo, S. T. A., & Hossain, N.

IEEE/CVF International Conference on Computer Vision (ICCV) Workshops · 2025, pp. 2955–2964
CORE A* workshop

MIMIC: Multimodal Islamophobic Meme Identification and Classification

Mustakim, S. H., Islam, S. M., Muna, U. M., Chowdhury, M., Islam, M. J., Ahmmed, S., Sikder, T., Dhrubo, S. T. A., & Shatabda, S.

3rd Muslims in ML Workshop — NeurIPS · 2024

View the full list on Google Scholar

Selected projects

Built end to end.

NurtureAid

A real-time, AI-assisted mobile application that simplifies elder and dependent care through caretakers. I led both backend and frontend as the full-stack engineer.

React Native
Node.js
Flask
MongoDB
Firebase
LLaMA-Index
PyTorch

Champion — UIU CSE Project Show, Software Engineering Laboratory
Champion — Hult Prize OnCampus, UIU
Selected for the Hult Prize Summit, Boston

Repository

Haven

A BCI-powered mental-health agent: speech to speech, reading emotion from the signal and answering in kind. I led the backend, frontend and AI work.

Python
OpenAI
Emotiv
Firebase
Streamlit

Winner — Harvard HSIL Hackathon, Dhaka Hub

Repository

Toolkit

What I reach for.

Programming

C
C++
Java
Python
JavaScript
PHP
SQL
Solidity

Machine learning

PyTorch
TensorFlow
Scikit-Learn
NumPy
Pandas
Matplotlib
NLTK
Hugging Face

Frameworks

React.js
Node.js
Spring Boot
FastAPI
Firebase

DevOps

Git
GitHub
Docker

And

Communication
Leadership
Design thinking
Public speaking

Achievements

A decent trophy shelf.

Winner — Harvard Health Systems Innovation Lab Hackathon, Dhaka Hub2025
Best Performing Engineer — SD12 Project of the Government of Bangladesh, given by Apurba Technologies Ltd.2025
Academic Excellence Scholarship, United International University2020 — 2024
Champion — Bhashamul: Bengali Regional Text to IPA National NLP Datathon2024
Champion — Hult Prize OnCampus Round2024
Summit Participant — Hult Prize Summit, Boston, USA2024
Gold Award (Champion) — International Blockchain Olympiad (IBCOL)2023
Silver Award (1st Runner-Up) — Bangladesh Blockchain Olympiad (BCOLBD)2023
1st Runner-Up — Intra-University Deep Learning Sprint2023
Finalist — ICT Innovation Grant, for NurtureAid2023
Champion — Software Engineering Lab2023
Champion — Database Management Systems Lab2022
Champion — Advanced Object-Oriented Programming Lab2022

Beyond the work

Teaching, speaking, organising.

General Secretary, UIU Computer Club

Ran the university’s computer club through two academic years.
2022 — 2024
Interview — Jamuna TV

On the triumph of IBCOL 2023 and the future of blockchain in Bangladesh.
Watch
Trainer — Automatic Speech Recognition workshop

Hands-on session with Bengali.AI and the IUT Computer Society.
Details
Trainer — Building a dApp from scratch

Hands-on workshop for the UIU Computer Club.
Details

Contact

Let’s talk.

Open to research collaborations in hierarchical and multi-modal learning, and to conversations about putting models in front of real users.

jishanlion@gmail.com

S M Jishanul Islam

Research that leaves the lab.

Where I learned it.

Master’s in Computer Science

B.Sc. in Computer Science & Engineering

Where I’ve built things.

AI Engineer

Part-time Faculty

Software Engineer

Lecturer (contractual)

Undergraduate Assistant

Answering one question.

Journals

An audio video-based multi-modal fusion approach for speech emotion recognition

Conferences

BornoDrishti: Leveraging Vision Encoders and Domain-Adaptive Learning for Bangla OCR on Diverse Documents

Beyond Characters: Position-Aware Metric and a Lightweight LLM for Low-Resource Bangla Punctuation Restoration

IPBlocks: A Blockchain Ecosystem for Secure IP Registration and Decentralized Marketplace

E-MedViTR: Enhanced Vision Transformers with Registers for Biomedical Image Classification

Workshops

Watch, Listen, Understand, Mislead: Tri-modal Adversarial Attacks on Short Videos for Content Appropriateness Evaluation

MIMIC: Multimodal Islamophobic Meme Identification and Classification