Software Development Engineer
This study explores sentiment analysis on Bitcoin-related tweets using two advanced neural network models: LSTM with attention and a Transformer model. Employing distinct tokenization methods, the Enhanced LSTM model focuses on model-specific features, while the Transformer leverages a pre-trained BERT tokenizer to handle complex linguistic patterns.
This project examines the performance of Q-learning and Deep Q-learning agents in the classic game of Pong. The exploration aims to gain deeper insights into the distinct ways reinforcement learning algorithms function in a gaming environment. By comparing the interactions between Q-learning and Deep Q-learning agents in Pong, this study seeks to understand their strengths and limitations. Rigorous experimentation across varying levels of complexity offers a nuanced perspective on how these algorithms adapt to dynamic gaming environments.
The Goal Tracker App is a robust and modular application designed to efficiently manage and track user goals. Its architecture includes distinct components such as templates that dynamically generate HTML content for the user interface, views that act as a bridge between the frontend and backend, and models that define the data structures and handle interactions with a PostgreSQL database. The backend is further reinforced by a user authentication module for secure access and a goals management module that handles the creation, updating, and tracking of goals.
A 2D game developed using Unity Engine and C#. Players can move (W, A, S, D), attack (Left Mouse Button), and collect props by walking over them. The goal is to survive against infinitely spawning enemies, with the final score being recorded upon the player's death.
A React project where users can add, delete, search, and comment on available places.
A program that uses a binary heap implemented in C++ to manage a hub airport. Flights arrive in a specific order but may have different departure times if delays occur.