Skip to content
View bayantuffaha's full-sized avatar

Block or report bayantuffaha

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

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

Report abuse
bayantuffaha/README.md

Hi, I’m Bayan

I’m an MS Data Science student at Tufts University with a BS in Applied Mathematics and Computer Science. Through coursework and internships, I’ve worked on projects spanning machine learning, probabilistic modeling, reinforcement learning, and full-stack web development.

I use this GitHub to share a small number of polished, representative projects.

Featured projects

  • Probabilistic Chess Outcome Prediction
    Multiclass logistic regression modeling implemented in Python, comparing MAP estimation via gradient descent and posterior predictive estimation via MCMC (NumPy, SciPy, PyMC).

  • Human Feedback in Reinforcement Learning (TAMER)
    Reinforcement learning experiments implemented in Python, evaluating event-based and navigation-guided feedback strategies in Cliff Walking and Taxi environments (Gymnasium, NumPy, matplotlib).

Additional project experience

  • Sentiment Analysis of Product Reviews — Machine learning classification model implemented in Python to predict review sentiment using BERT embeddings and multilayer neural networks (scikit-learn, pandas, NumPy).
  • Movie Recommendation System — Collaborative filtering model implemented in Python to predict user ratings, trained with stochastic gradient descent and evaluated using MAE and RMSE (Surprise, scikit-learn).
  • Design&Dine Web App — Full-stack web application for user-tailored recipe planning using third-party APIs, server-side logic, and a relational database, with features for saving favorite recipes, managing a shopping list, and ensuring user-friendly design (PHP, MySQL, JavaScript, HTML, CSS).

Some projects are not public due to academic integrity policies or because they are still being curated for release; I’m happy to discuss additional work upon request.

Pinned Loading

  1. predict-chess-outcomes predict-chess-outcomes Public

    Probabilistic multiclass prediction of chess outcomes using logistic regression, comparing MAP estimation and posterior predictive estimation via MCMC

    Jupyter Notebook

  2. human-feedback-rl human-feedback-rl Public

    Experimental study of human feedback strategies in reinforcement learning using the TAMER framework across classic navigation environments

    Python