Projects

A curated portfolio of research projects in medical imaging, information retrieval, and generative deep learning, emphasizing reproducibility, generalization, and real-world evaluation.

work

project thumbnail

Skin Lesion Classification with Deep Learning Ensembles

Medical Imaging | Melanoma Detection | Explainable AI

A generalizable ensemble deep learning framework for automated skin lesion classification, integrating CNNs and ViTs with internal + external validation.

project thumbnail

Bangladesh AI Policy & Innovation Lab (BAPIL)

AI Governance | Policy Intelligence | Digital Government

An independent research initiative proposing an AI-powered policy intelligence framework for evidence-based policymaking, responsible AI governance, and digital transformation in Bangladesh.

project thumbnail

Patent Retrieval & Re-ranking (Dense + Cross-Encoders)

Information Retrieval | Dense Retrieval | Cross-Encoder Re-ranking

Research-oriented patent retrieval pipeline combining dense retrieval with transformer cross-encoder re-ranking, evaluated using standard IR metrics.

project thumbnail

CycleGAN: Unpaired Horse-to-Zebra Image Translation

Generative Adversarial Networks | Unpaired Image Translation | Computer Vision

PyTorch implementation of CycleGAN for unpaired image-to-image translation, with qualitative results and quantitative evaluation using SSIM and PSNR.

project thumbnail

Emotion Recognition from Text (ML + Transformers)

Natural Language Processing | Emotion Classification | Transformers

Comprehensive emotion recognition pipeline combining classical machine learning baselines with transformer-based models, evaluated on a standard emotion dataset.

project thumbnail

Simplified DRAW Model for Generative Image Modeling

Generative Modeling | Variational Autoencoders | Representation Learning

Implementation of a simplified DRAW-style recurrent variational autoencoder for iterative image reconstruction and generation on Fashion-MNIST.