My Projects
-
Implementation of various machine learning and deep learning research papers, covering recent advancements in the field.
Python Numpy PyTorch TensorFlow -
Conducted a sentiment analysis on BinusMaya, Binus University's LMS, using feedback from 600+ students. Employed Naive Bayes and K-Nearest Neighbors (KNN) with cross-validation for robust classification and model evaluation. Provided key insights into user experience and platform satisfaction.
Python scikit-learn NLTK -
Developed an Emergency Autoresponder System that uses existing CCTV infrastructure to detect traffic accidents and automatically alert nearby medical and police personnel. Built accident detection system using YOLOv5 deep learning models with mAP 0.50 and mAP 0.50-0.95 validation metrics. Created intuitive GUI interfaces using tkinter and successfully deployed functional prototype with high detection accuracy.
Python YOLOv5 OpenCV tkinter -
Developed an image captioning model using PyTorch, combining a ResNet152 for visual feature extraction with attention block and LSTM decoder. Trained and evaluated on the Flickr8k dataset.
Python PyTorch Transformers Attention -
Built an image classification model using a custom CNN and enhanced performance through transfer learning with DenseNet121.
Python TensorFlow Keras -
Developed a deep learning model using LSTM to classify news articles into categories like politics, sports, and technology.
Python TensorFlow Keras -
A deep learning model using Convolutional Neural Networks to detect Parkinson's disease from medical imaging data.
Python TensorFlow Keras -
Created a data visualization project analyzing online gambling trends in Indonesia (2017–Q1 2024), highlighting demographic patterns, transaction frequency, and financial flow over time.
Python Matplotlib Pandas -
Machine learning model to predict used car prices in the Indonesian market based on various vehicle attributes.
Python scikit-learn Pandas Streamlit -
Book recommendation engine using large language models (LLMs) to suggest books based on content themes, writing styles, and emotional tones, by analyzing book descriptions.
Python Gradio LLM -
This project is part of my fun exploration into web development. It's a playful web app that turns URLs or text into QR codes, all while embracing a retro Linux terminal aesthetic. A simple and enjoyable project to showcase QR code generation with a stylish twist!
HTML CSS JavaScript