Tanjim Islam Riju

About Me

I work at the intersection of computer vision and multimodal machine learning, with a focus on research that is reproducible, evidence grounded, and rigorously evaluated. I care a lot about understanding failure modes, hallucinations, shortcut learning, and brittle generalization, especially in high-stakes settings like medical imaging.

I’ve worked on gaze-supervised learning for chest X-rays, using human attention signals to better align diagnosis and support more faithful report generation. I’ve also contributed to long-document and long-context evaluation as part of a broader goal: making model limitations visible and measurable through stronger benchmarks and analysis.

Alongside research, I bring solid software engineering experience. I build end-to-end systems and research tooling, often with React and Tailwind on the frontend, Go on the backend, plus Docker and AWS for practical deployment.

I’m preparing to pursue a PhD in AI/ML. My goal is to advance reliable computer vision and multimodal methods that are both scientifically grounded and genuinely useful.

Education

B.Sc. in Computer Science and Engineering

BRAC University•Graduated with Distinction

Thesis: Medical Image Analysis

Relevant Coursework: Artificial Intelligence (CSE422), Neural Networks (CSE425), Algorithms (CSE221), Data Structures (CSE220), Discrete Mathematics (CSE230), Computer Graphics (CSE423).

Jan 2020 – April 2024

Experience

Invicta Solutions Limited

June 2024 – Present

Fullstack Developer

Developed scalable web and mobile applications using React.js, React Native, and Node.js.
Built and deployed containerized backend services with AWS ECS and Docker, improving system reliability by 35%.
Designed visual elements and UI mockups using Figma and Adobe XD.
Collaborated with design and backend teams to integrate REST APIs and assets seamlessly.

Kaz Software Ltd.

July 2023 – May 2024

Software Engineer

Designed RESTful services in Node.js with PostgreSQL, increasing data retrieval efficiency by 25%.
Ensured UI consistency across projects using TailwindCSS and Adobe Illustrator, reducing inconsistencies by 40%.
Migrated microservices to AWS Lambda, improving reliability by 30% and cutting operational costs by 20%.

Infolytx

Jan 2023 – June 2023

Backend Engineer

Developed REST APIs and backend pipelines using Python and PostgreSQL.
Built internal dashboards with custom graphics using Photoshop and Dreamweaver.

LeadSoft Bangladesh Ltd.

July 2022 – Dec 2022

Mobile Developer Intern

Built mobile UI components and integrated Firebase for real-time user tracking.

DataSoft Systems

Jan 2022 – June 2022

DevOps Intern

Automated CI/CD pipelines and managed containerized deployments using Docker and Linux servers.

Genex Infosys Ltd.

Sep 2021 – Dec 2021

UI/UX Design Asst.

Designed web and mobile graphics using Photoshop, Illustrator, and Dreamweaver.

Research

Eyes on the Image: Gaze Supervised Multimodal Learning for Chest X-ray Diagnosis and Report Generation

Tanjim Islam Riju, Shuchismita Anwar, Saman Sarker Joy, Farig Sadeque, Swakkhar Shatabda

Preprint

arXiv:2508.13068

Focused on leveraging gaze data and multimodal contrastive learning to improve medical AI systems. Developed frameworks integrating vision-language models for diagnosis and report generation.

Gaze SupervisionMultimodal LearningMedical AIChest X-rayVision-Language Models

Projects

ResNet18 Pneumonia Classification (PneumoniaMNIST)

Implemented ResNet18 for pneumonia detection from chest X-rays.

Python, PyTorch, Medical Imaging

VGG19 Ultrasound Classification

Implemented VGG19 in TensorFlow/Keras for ultrasound image classification.

Python, TensorFlow, Keras, Computer Vision

SENet Ultrasound Classification

Applied Squeeze-and-Excitation Networks (SENet) for ultrasound image recognition.

Python, Deep Learning, Medical Imaging

Breast Cancer Prediction Using Machine Learning

Explored multiple ML algorithms for breast tumor classification (benign vs malignant).

Python, Scikit-learn, Machine Learning

Kidney Disease Prediction

Evaluated Logistic Regression, Random Forest, SVM, XGBoost, Gradient Boosting for CKD detection.

Python, XGBoost, SVM, Random Forest

Liver Disease Prediction

Applied multiple ML models to classify liver disease likelihood.

Python, Machine Learning, Healthcare

Heart Disease Prediction

Compared classical ML models for cardiovascular risk detection.

Python, Logistic Regression, Decision Trees, SVM

Diabetes Prediction

Built multi-model system with Logistic Regression, KNN, SVM, RF, Gradient Boosting, XGBoost.

Python, XGBoost, Ensemble Methods

Parkinson's Disease Prediction Using SVM

Developed SVM-based biomedical model for Parkinson's disease detection.

Python, SVM, Biomedical Data

Emotion Prediction Using Neural Networks

Classified emotions from text using embedding layers and dense neural nets.

Python, Neural Networks, NLP

Fake News Detection with FaKnow

Leveraged PyTorch-based FaKnow library to detect fake news using content and social signals.

Python, PyTorch, NLP, Social Media

Book Recommendation System

Built recommendation engine using TF-IDF and cosine similarity.

Python, TF-IDF, Recommendation Systems

Fraud Detection Using Decision Tree Classifier

Designed ML model for detecting fraudulent transactions.

Python, Decision Trees, Fraud Detection