Machine Learning · Deep Learning · Natural Language Processing
I'm a final-year BS Computer Science student at COMSATS University Islamabad, Lahore Campus. I build practical AI/ML projects and explore how models work, how to evaluate them, and how to turn them into usable applications.
My interests include low-resource NLP, Urdu language technology, computer vision, and retrieval-augmented generation.
| Project | Focus | Main tools |
|---|---|---|
| CodeSage | A codebase question-answering assistant with retrieval and file citations | LangChain, Chroma, Streamlit |
| LinguaNext | LSTM next-word prediction for English and Roman Urdu | PyTorch, FastAPI, React |
| Weather Image Classifier | Comparing a CNN trained from scratch with a fine-tuned ResNet34 | PyTorch, FastAPI, Streamlit |
| Concrete Strength Predictor | Predicting concrete compressive strength from its input properties | Python, machine learning |
| DocuLegal | PDF question answering with hybrid retrieval and source citations | LangChain, Chroma, Streamlit |
| Email Spam Classifier | Spam classification using TF-IDF and classical ML | scikit-learn, Streamlit |
HateShield — Urdu and Roman Urdu hate-speech detection
My final-year research project explores multi-task hate-speech and abusive-language detection across Native Urdu and Roman Urdu. My work includes Native Urdu data collection and dataset preparation. The project is ongoing, with an emphasis on annotation quality and careful evaluation.
- ML and deep learning: Python, scikit-learn, PyTorch, pandas, NumPy
- Applications: FastAPI, Streamlit, React
- Retrieval and NLP applications: LangChain, Chroma, Hugging Face models
I completed an AI/ML internship at AllZone Technologies, where I learned machine learning and deep learning fundamentals and built projects. I have also completed Andrew Ng's Machine Learning Specialization.
I'm continuing to develop my skills through university coursework, practical projects, and research.
Contact: muhammadtaimoor1407@gmail.com