Skip to content
View najmulhasan-code's full-sized avatar

Block or report najmulhasan-code

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
najmulhasan-code/README.md

Najmul Hasan

GitHub LinkedIn Twitter

Hello! I am a senior at UNC Pembroke pursuing a BS in Computer Science with minors in Mathematics and Physics.

I focus on using reinforcement learning (RL) to improve reasoning in large language models (LLMs). I work with Dr. Shaohu Zhang (NC A&T, prev. UNCP) on developing better reward design for LLM reasoning. Currently, I am developing step level intrinsic calibration (SLIC), which combines process supervision with confidence measurement to train models that are both accurate and well calibrated about their uncertainty.

I also work with Dr. Prashanth BusiReddyGari on drone security applications using LLMs. Previously, I have worked on multilingual phishing email detection with LLM with Dr. BusiReddyGari and Dr. Zhang, speech emotion recognition with Dr. Zhang, phishing URL detection with LLM with Dr. BusiReddyGari, and lightweight cryptography algorithms and mobile driver's license (mDL) with Dr. BusiReddyGari and Dr. Ali Saman Tosun.

I'm applying to PhD programs for Fall 2026. I'm always happy to discuss research ideas or potential collaborations. Feel free to reach out!

Pinned Loading

  1. SummaryOne SummaryOne Public

    Modern AI-powered text processing platform with customizable summarization, translation, grammar checking, content expansion, and tone adjustment features.

    TypeScript 4

  2. text-sentiment-analyzer text-sentiment-analyzer Public

    A web application for analyzing the sentiment of text using a trained machine learning model. Tech Stack: Sentiment140 dataset with 1.6 million tweets, Naive Bayes classifier, Python, Skit-learn, N…

    Jupyter Notebook

  3. using-gemma-to-answer-common-python-questions using-gemma-to-answer-common-python-questions Public

    Using Gemma to Answer Common Python Questions | Python, Gemma-2b-it Independent Project/Kaggle Competition Entry

    Jupyter Notebook

  4. -Concrete-Strength-Prediction-using-Neural-Networks -Concrete-Strength-Prediction-using-Neural-Networks Public

    Developed a Keras-based neural network model to estimate concrete strength. Implemented data normalization and rigorous model evaluation. Executed 50 training iterations to assess model stability, …

    Jupyter Notebook