Artistic Neural Style Transfer

Technologies: PyTorch, Jupyter Notebooks, Streamlit | Oct 2024 - Present
GitHub Repository

  • Implemented Neural Style Transfer using VGG19 CNN for feature extraction and gram matrix computations
  • Developed Streamlit app enabling users to generate stylized images interactively
  • Fine-tuned hyperparameters to optimize style transfer quality

Neural Transfer Diagram illustrating the Neural Style Transfer process using content and style images

Streamlit Interface Interactive web interface for style transfer

Accent Detection Deep Learning Model

Technologies: PyTorch, Pandas, Scikit-Learn, Librosa | Apr 2024 - Present
GitHub Repository

  • Reviewed literature on accent recognition methods and deep learning architectures
  • Collected audio recordings from Speech Accent Archive
  • Preprocessed dataset using silence trimming and noise reduction
  • Implemented and fine-tuned ANN to predict accents using MFCCs
  • Achieved: 0.84 accuracy, 0.79 F1 score, 0.80 precision

Student Accommodation Clustering

Technologies: Python, Scikit-Learn, Folium | Nov 2022 - Dec 2022
GitHub Repository

  • Developed clustering system to suggest optimal housing options based on proximity to preferred locations
  • Conducted data collection and cleaning, including API integration and handling missing values
  • Implemented K-Means clustering algorithm from scikit-learn to group accommodation options
  • Performed exploratory data analysis with visualizations and interactive maps

Map Clustering

Optimal housing recommendations based on students’ preferred locations

Financial Cointegration Analysis

Technologies: Time Series Analysis, Python, Jupyter Notebooks | Dec 2023
GitHub Repository

  • Applied Engle-Granger two-step method to distinguish spurious correlations from genuine long-term relationships
  • Analyzed historical price data from Yahoo Finance using yfinance library
  • Demonstrated cointegration testing methodology for pairs trading strategies

Cointegration Testing Testing cointegration between two financial time series