In this project, we implemented and extended the research paper. The work was completed in two phases: first, reproducing the original methodology and results; and second, enhancing the model using attention mechanisms and dynamic threshold optimization.

In this project, we built a full-stack application that fetches trending topics from Google Trends, generates creative stories using a language model, and optionally converts them to audio. It features genre customization, a web interface, and a gRPC-based microservice architecture.

This project implements an end-to-end traffic sign classifier that operates entirely using classical image processing techniques.

• Developed a 2D shooting game in Assembly, featuring Duck as the target, with real-time input handling and dynamic difficulty scaling.

This project aims to predict flight departure delays using raw datasets, including weather and flight data. By analyzing delay patterns and building predictive models, it generates accurate predictions for a Kaggle competition.

The project involves predicting the age of speakers from audio files by applying a linear regression model. The approach includes custom implementations of linear regression and various preprocessing techniques, followed by comparisons with scikit-learn's built-in models.

This project aims to generate an optimal exam timetable for a set of courses and students using a genetic algorithm. The algorithm takes into account both hard and soft constraints to ensure the most efficient scheduling while maintaining fairness for students and teachers.

In this project, we implemented a basic MapReduce framework using core OS concepts. It simulates distributed data processing on a single machine and demonstrates the key phases of MapReduce: Map, Shuffle, and Reduce.

In this project, I developed EventKhaana, a platform that showcases a wide range of verified halal events across Pakistan. It supports both event organizers and participants, aiming to foster meaningful connections and memorable experiences.