I am Arnab Bhattacharjee, currently pursuing my BSc degree in Computer Science & Engineering at Bangladesh University of Engineering and Technology (BUET).

I’m currently exploring Large Language Models (LLMs), Machine Learning and Web Development. I enjoy solving problems, building impactful solutions, and continuously learning new technologies.

(Click on the project name to view the repository)

• Technologies: React.js, SpringBoot, PostgreSQL, Tailwind CSS, Firebase (Auth)
• Description: A multi-purpose platform for engineers to showcase, monetize, and enhance their skills.
  • Learning: AI Chatbot, smart notes app, and AI tools for summarizing and paraphrasing text.
  • Showcasing: Create posts and display public projects for better visibility.
  • Monetizing: Marketplace for project sales, with order tracking and transaction management.
  • Jobs: Auto-generated portfolios and integrated public blogs.

• Technologies: React.js, FastAPI, PostgreSQL, Tailwind CSS, SQLAlchemy, Supabase (DB and Auth)
• Description: A platform designed for Bengali-speaking students featuring:
  • Knowledge Base Chat: Users can interact with uploaded knowledge bases (e.g., files) in Bangla, Banglish, or Audio.
  • Smart Note Editor: Create notes in Banglish, converted into proper Bangla using LLMs.
  • Model Enhancer: Continuously improves Banglish-to-Bangla translation models.

• Technologies: Node.js, Express.js, EJS, PostgreSQL, Oracle DB
• Description: An e-commerce website for selling PCs and accessories.
  • Key feature: PC-BUILDER, a custom PC-building tool.
  • Implements advanced database operations with SQL and PL/SQL programming.

• Technologies: Python, PyTorch, YOLOv11
• Description: A robust image detector for classifying real and fake images under adversarial attacks.
  • Black-Box Experiments: Evaluated models like ViT and YOLOv11x on noisy datasets, achieving up to 80% accuracy.
  • White-Box Experiments: Enhanced robustness using custom adversarial noise and dropout layers.

• Technologies: Python
• Description: Implemented a Feed-Forward Neural Network (FNN) for apparel classification with advanced features like:
  • Batch Normalization
  • Dropout
  • Adam Optimization

Additional assignments explored foundational concepts in Machine Learning.

Topic: Enabling Multilingual Interaction with User Interface

• Objective: Develop a multilingual web agent capable of performing complex multi-step tasks on web pages.
• Research Areas: LLM Agents, Human-Computer Interaction, UI Understanding, and Large Language Models (LLMs).

• 🥇 Champion: ITVerse 2023 ‘TechTales’ Project Showcase Competition at Dhaka University.
• 🏅 Finalist: BUET CSE Fest Capture the Flag (CTF) Competition 2023.

Feel free to explore my repositories and connect with me. Let’s collaborate on building impactful solutions! 😊