Angus Saw's Project Portfolio

Greetings! I am a data science practitioner with industry experience in data visualization projects and end-to-end machine learning projects. I am interested in building tools and programs using data science techniques to solve complex problems, improve decision making, automate processes and drive value creation.

Here is my portfolio showcasing some of my personal projects where I apply what I have learned throughout my career on topics that I find interesting!

Skills

  • Data Analysis and Visualization
  • Self Motivated
  • Feature Engineering
  • Detail Orientated
  • Model Evaluation and Experimentation
  • Communication and Collaboration

Personal projects

Conversing with multiple documents simultaneously via retrieval augmented generation and conversation agents

Developed an chatbot that enables users to converse with multiple documents at once via retrieval augmented generation from large language models

Estimating and interpreting HDB resale prices via feature engineering and SHAP values

Performed web scraping to obtain geographical coordinates to estimate HDB resale prices and enhance model explainability using SHAP values

Explainable-Neural-Networks for pneumonia detection using Gradient-weighted Class Activation Mapping (Grad-CAM)

Using a Grad-CAM algorithm to explain X-ray image classifications made by a deep learning model trained using a transfer learning framework