Projects | Atena Jafari Parsa

Cellenta iOS

An online charging system iOS app built with SwiftUI during my internship at i2i Systems. It integrates authentication, package management, and billing with a backend API. Equipped with a seamless live chat bot and AI-powered smart, personalized package recommendation system using Google Gemini's API. Developed with Swift.

View Project

Watch Project Demo

Aware Citizen

An AI-Based Political Economy Analyst web app built with Streamlit and empowered by Open AI's API that finds the political parties of the selected country with Beautiful Soup from Wikipedia, and uses Open AI's Turbo model to analyze the effects of every party on citizens' daily lives if selected. It assigns a score from 0 (very bad) to 20 (very good) to each party and visualizes a comparative result in the end. Developed with Python and Streamlit.

View Project

Research Paper

Project Presentation

Pima Diabetes Prediction

Machine learning pipeline using Pima Indians Diabetes Dataset on Kaggle for diabetes risk prediction using MLflow to track experiments, reproducibility, and results. A predictive ML model (classification task for diabetes) was implemented and its end-to-end lifecycle was managed, consisting of training, hyperparameter tuning, deployment, performance monitoring, and model versioning. 5 different models were trained and tested: Random forest, Support Vector Machine (SVM), KNeighbors Classifier, Decision Tree Classifier and Logistic Regression. Hyperparameters of the model reaching the highest accuracy was tuned for optimization. Developed with Python and utilized MLflow.

View Project

Project Report

Project Presentation

BAU-EVAL

AI-based Quiz and Exam Generator & Evaluator developed for Bahcesehir University as my Senior Capstone Project empowering instructors with: LLM-Generated Exam & Quiz Questions, Automated Grading with Instructor Approval, Personalized Smart Feedback for Students, Seamless Student and Grade Management.

Project Report

Project Presentation

Re [Style TTS2]

A reproduction study of Style TTS2: Towards Human-Level Text-to-Speech through Style Diffusion and Adversarial Training with Large Speech Language Models which utilizes deep learning and artificial neural networks.

View Git Repository

Original Paper

Stroke Prediction with Multiple Linear Regression

Stroke Prediction with Multiple Linear Regression based on the kaggle competition Stroke Prediction Dataset Assignment using machine learning with Python and Scikit Learn.

View Project

Project Report

Machine Learning and The Titanic Disaster

a machine learning model to predict whether a passenger survived or not during the tragic event of the Titanic disaster. Based on Titanic - Machine Learning from Disaster competition on Kaggle. Developed with Python and Scikit Learn.

View Project

Project Report

Ensemble Learning

Comparing the accuracy in an example problem by classifying by soft voting, hard voting, using bagging/pasting, Out of Bag Evaluations, Random Forests and my own novel ensemble model. Utilized soft voting classifiers On SVM (Soft Vector Machine), KNeighbor, Random Forest and the different combinations of them. Developed with Python and Scikit Learn.

View Project

Project Report

Advertising Effects on Sales with R

This data science project analyzes the impact of three advertising methods—Television (TV), Radio, and Newspaper—on sales performance. The dataset includes 200 observations of advertising budgets and corresponding sales figures, measured in thousands of dollars and units respectively. The goal is to determine which advertising methods significantly influence sales and to model these relationships using statistical techniques.

View Project

Project Report

Project Presentation

Insurance Premium Calculator

An OOP python app to create a simple system for an insurance company with 3 types of insurance policies; health policy, auto policy, and life policy. This project aims to create a simple system for an insurance company ensuring that the information of the customers are stored privately, and each policy is calculated according to different criteria of the customers’ information. Developed with Python.

View Project

Project Report

Library Web Application

An abstract library web application developed with SpringBoot (SpringToolSuite4) using an h2-database that represents functionalities from the users and the admin as well as performing CRUD operations on the books. Developed with SpringBoot in Java.

View Project

Project Report

Project Presentation

Hazelcast Demo Sales Project

A Java-based demo application that simulates distributed order processing using Hazelcast. Developed during my internship at i2i Systems, the project demonstrates how to set up a single-node Hazelcast cluster, use IMap for in-memory data storage, and perform basic get/put operations. It also serves as an introductory example for understanding distributed caching and partitioning in real-time systems. Developed with Java.

View Project

Project Presentation

Book Comet Web Application

Developed a library web application using Flask that allows users to manage a bookstore by adding, updating, and deleting books and integrated Azure Cosmos DB to handle backend data storage, ensuring scalable and efficient management of book information. Deployed and tested the application locally, providing clear instructions for setting up the environment and running the app. Developed with Python and deployed via Azure Kubernetes.

View Project

Project Report

The Unsaid Tales

An interactive console-based story-telling program written in Java. This project allows users to explore a curated collection of genres and stories. The program offers a menu-driven interface, providing options to either select a story from a specific genre or play a randomly selected story. With a visually engaging display and dynamic interaction, it delivers an immersive storytelling experience. Written in Java on Ready to Program.

View Project