AI Researcher · Competitive Programmer · CSE Student
I build research driven AI systems using machine learning, deep learning, large language models, RAG, explainable AI, and real time intelligent applications.
$ profile
name: Md. Tanvir Mahamud Himel
focus: ML, XAI, LLM, RAG, CV, EEG
status: final year CSE student
github: github.com/404himel
// About Me
I am a final year B.Sc. student in Computer Science and Engineering at Comilla University. My work focuses on machine learning, deep learning, large language models, explainable AI, real time intelligent systems, and data driven research.
I have built systems for sign language recognition, diabetes prediction with explainable AI, EEG based attention decoding, RAG based question answering, AI health assistants, and speaking evaluation tools.
// Projects
Fine tuned LLaMA 3 8B with 4 bit quantization using Unsloth and connected it with SQLite based natural language to SQL for real time banking support.
An AI speaking simulator with examiner voice, speech transcription, band scoring, and feedback for IELTS speaking practice.
Generated personalized diet and exercise plans using RAG, real time LLM responses, sentence embeddings, ChromaDB, and automated PDF reports.
Created a document grounded question and answer generator with RAG, LLaMA 3.1 8B through Groq, LangChain, embeddings, and CSV export.
Built a Selenium based scraper to collect product information, prices, and reviews with multithreading for faster data collection.
// Skills
TensorFlow, PyTorch, Scikit learn, Keras, CNN, RNN, LSTM, Transformer, ViT, Ensemble Methods, SHAP, LIME
Hugging Face Transformers, LangChain, LLaMA 3, RAG, Sentence Transformers, fine tuning, Unsloth, NL to SQL
NumPy, Pandas, Matplotlib, Seaborn, preprocessing, feature engineering, statistics, probability
MySQL, SQLite, ChromaDB, vector database based retrieval systems
Git, GitHub, Google Colab, Jupyter Notebook, VS Code, Overleaf, Streamlit, Django, Docker basics
C, C++, Python, Java, basic JavaScript, problem solving, algorithmic thinking
// Research
Vision Transformer based real time sign language interpretation system, presented at the 27th International Conference on Computer and Information Technology.
A stacking ensemble approach for better accuracy and interpretability in diabetes prediction.
Accepted at IEEE Computer Society Bangladesh Chapter Summer Symposium 2025.
An intelligent robotic system for real time attention level monitoring, under review in Engineering Applications of Artificial Intelligence.
Attention aware Transformer architecture for decoding human attention levels from EEG signals.
A lightweight Transformer based efficient sentiment classification system for Bangla e commerce reviews.
// Experience
Conducting collaborative international research in artificial intelligence and machine learning.
Working on HCI based student attention modeling and Transformer based EEG signal attention decoding.
Studying computer science and engineering with a current CGPA of 3.70 out of 4.00.
Ranked 412th out of 3000 plus teams in ICPC Asia Dhaka Regional Preliminary and solved 600 plus problems.