About Me

I am an AI/ML researcher and graduate student specializing in multimodal AI, computer vision, and natural language processing. My work focuses on practical applications in medical imaging and computational methods for under-resourced languages.

I am currently a Research Assistant at the North South University, also contributing to the Centre for UnlieAl Artificial Intelligence Research Lab . My research involves dataset preprocessing, model training, and preparing publications for premier AI conferences, with papers currently under review at venues including NeurIPS and ACL

My technical expertise includes building and training deep learning models using PyTorch, and Hugging Face Transformers. I have extensive experience with Vision-Language Models (VLMs) such as CLIP, VILT, BLIP, and BioViL-T for developing radiology benchmarks and naunce hate detection.

Skills & Expertise

AI & Machine Learning

PyTorch TensorFlow Hugging Face Transformers Scikit-learn XGBoost LightGBM

Domains & Applications

Multimodal AI Computer Vision Natural Language Processing (NLP) Medical Imaging (Radiology) Time Series Forecasting Under-resourced Languages

News

  • Feb 2025: Started role as Lab Instructor at North South University.

  • Jan 2025: Joined the Centre for UnlieAl Artificial Intelligence Research Lab at CU Boulder as a Research Assistant.

  • Oct 2024: Our paper, "RadMultiBench," was submitted to ICDE 2026.

  • Oct 2024: Our paper on "Consensus Feature Selection Pipeline" was submitted to ICDE 2026.

  • Apr 2024: Graduated from North South University with a B.S. in Computer Science and Engineering (Summa Cum Laude).

Publications

Preprints & Under Review

RadMultiBench: A Multimodal Benchmark Reveals Architectural Failures in Modern Vision-Language Models for Radiology

Md. Nahin Alam, et al.

Submitted to ICDE 2026

Paper Code

Live Apps

AI • Video • LLM

Chronicle Creators 🎬

An AI-powered faceless video generation platform that transforms any topic into a complete, engaging video using LLM-based script generation, text-to-speech narration, timed captions, and automated background video retrieval. Built with Python 3.11 and Gradio for an interactive web experience.

  • • End-to-end AI video automation pipeline
  • • LLM integration for dynamic script generation (Groq API)
  • • Text-to-Speech (TTS) audio synthesis with caption alignment
  • • Stock video search & matching via Pexels & Pixabay APIs
  • • Deployed as a live Hugging Face Space
Python Gradio LLMs AI Video Generation APIs Hugging Face
View Project
● Live

RBAC & Project Management System

An enterprise-grade administration platform engineered on the PERN Stack (PostgreSQL, Express.js, React 18, Node.js). It implements a strict MVC Architecture with TypeScript for type safety across the full stack. Security features include granular Role-Based Access Control (RBAC), cryptographically secure invitation workflows, and JWT stateless authentication. The backend leverages Prisma ORM for complex relation handling and Soft Delete mechanisms to ensure data integrity. The frontend utilizes Redux Toolkit for deterministic state management and Axios Interceptors for seamless token injection, wrapped in a responsive Material UI (MUI) interface.

TypeScript React 18 Redux Toolkit Node.js Express.js PostgreSQL Prisma ORM REST API MUI JWT & Bcrypt Render/Vercel
View Project
● Live

AntHive: Real-time Language Exchange Platform

A comprehensive social application built on the MERN Stack (MongoDB, Express.js, React 19, Node.js). It features high-performance real-time messaging powered by Stream Chat and seamless video conferencing integrated via the Stream Video SDK. State management is handled efficiently with Zustand, while server-side state and caching are optimized using TanStack Query. The UI is responsive and themeable, crafted with Tailwind CSS and DaisyUI. Security is enforced via JWT Authentication and HTTP-only cookies.

MERN Stack React 19 Vite Stream Chat SDK Stream Video SDK Zustand TanStack Query MongoDB Express.js Tailwind CSS DaisyUI JWT Auth
View Project
● Live

CodeGenius - AI Learning Platform

A personalized coding education platform built with Next.js 15 and React 19. It leverages Google Genkit and the Gemini 2.5 Flash model to generate custom learning paths and real-time code examples. The backend is fully serverless using Firebase (Auth, Firestore, App Hosting) for seamless scalability. Features include an interactive dashboard, secure authentication, and a polished interface styled with Tailwind CSS and Shadcn UI.

React 19 Next.js 15 TypeScript Google Genkit Gemini 2.5 Flash Firebase Tailwind CSS Shadcn UI Zod Validation Lucide Icons App Hosting
View Project
● Live

AI-Powered Resume Builder

A full-stack SaaS application that leverages Generative AI to automate resume creation. Built with React and Vite for a high-performance frontend, it integrates Google Gemini AI to generate professional summaries and experience points contextually. The backend is powered by a Strapi Headless CMS providing robust content management and **REST APIs**. Features include Clerk Authentication for secure user sessions, real-time live preview, one-click PDF export, and social sharing capabilities. Styled with Tailwind CSS and Shadcn UI for a modern, responsive aesthetic.

React 18 Vite Full Stack Google Gemini AI Strapi CMS REST API Tailwind CSS Shadcn UI Clerk Auth Real-time Preview Vercel Deployment
Try Live App
● Live

Duari: AI Study Abroad Companion

A full-stack education platform engineered with Next.js 15 (App Router) and TypeScript. It leverages Google's Genkit and the Gemini 2.5 Flash model to generate personalized university recommendations. The system features a robust backend using Firebase for auth/data and Prisma ORM, interactive UI with Shadcn UI and Tailwind CSS, and data visualization via Recharts. Includes drag-and-drop tracking (DnD Kit) and strict schema validation with Zod.

Gemini 2.5 Flash Genkit AI Next.js 15 React 18 TypeScript Firebase Prisma ORM Tailwind CSS Shadcn UI Zod React Hook Form Recharts DnD Kit Radix UI
View Project
● Live

Nexus: AI Knowledge Cartographer

An interactive research assistant that visualizes complex topics as a dynamic knowledge graph using the Gemini 2.5 Flash model. Built with a React/TypeScript frontend, it features a real-time chat interface to initiate graph generation. The application leverages the Gemini API's Structured Output (JSON) capability to precisely define Knowledge Nodes and Edges, which are rendered in an interactive SVG visualizer.

Gemini 2.5 Flash React & TypeScript Knowledge Graph Generation Structured Output (JSON) Tailwind CSS & Vite
View Project
● Live

CodePrep | Interview Platform

A comprehensive Full-Stack algorithm practice platform engineered with Next.js 14 (App Router) and TypeScript. Features a secure Remote Code Execution sandbox, robust authentication via Supabase Auth, and persistent data storage using PostgreSQL. Implements Row Level Security (RLS) for data protection, Server Actions for mutation, and a responsive UI built with Tailwind CSS and Shadcn UI.

Next.js 14 TypeScript Supabase & PostgreSQL Server Actions Tailwind & Shadcn
View Project
● Live

Real-Time Chat Platform

A full-stack messaging application built with the MERN stack. Features real-time communication via Socket.io, secure JWT authentication, global state management with Zustand, and live user status updates.

MERN Stack Socket.io Tailwind & DaisyUI JWT
Launch App
● Live

Full-Stack E-Commerce App

A production-ready platform engineered with the MERN stack and Ant Design. Features a comprehensive RESTful API, secure JWT authentication, admin dashboard for inventory management, and integrated Braintree payment processing.

MERN Stack Ant Design Braintree Payments Context API
Visit Shop

Experience

Research Assistant (remote)

Centre for UntieAI Artificial Intelligence Research Lab

Jan 2025 – Current

Conducting research in multimodal AI for medical imaging and NLP for under-resourced languages.

Technical Specialist

The Data Island, Dhaka

Jun 2024 – Oct 2025

Architected and deployed scalable solutions using AWS, GCP, Docker, and Kubernetes. Implemented CI/CD pipelines.

University Research Assistant

Machine Learning Lab, North South University

Jan 2025 – Oct 2025

Led research on multimodal AI benchmarks (RadMultiBench) and developed feature selection pipelines for biomarker discovery.

Research Assistant

Cyber Physical Systems Lab, North South University

Jun 2023 – Jun 2024

Developed intelligent sensor networks and implemented ML algorithms for predictive maintenance in IoT systems.

Teaching

Lab Instructor

North South University

Feb 2025 – Current

Teaching and conducting labs for AI and Machine Learning courses. Designing exercises using PyTorch, TensorFlow, and scikit-learn.

University Teaching Assistant

North South University

Jun 2024 – Jan 2025

Assisted faculty with undergraduate courses. Supervised labs for programming (Python, Java, C/C++) and held sessions on algorithms and data structures.

Education & Certifications

Education

Bachelor of Science, Computer Science and Engineering

North South University, Dhaka

Jan 2020 – Apr 2024

  • Final Grade: 3.81/4.00 (Summa Cum Laude)
  • Honors: Dean's List recognition for academic excellence

Certifications

  • FastAI Deep Learning Course
  • Stanford Machine Learning Specialization (Coursera)
  • Andrew Ng Deep Learning Specialization (Coursera)
  • freeCodeCamp Data Science Certification
  • Google Cloud Platform Fundamentals