Ahon

Welcome to my portfolio! Explore my work, academics, and how to contact me.

About Me

Hi, I'm Ahon, a second-year Physics student at Shiv Nadar University, Greater Noida. I'm interested in machine learning, data analysis, and graphics programming. I've been programming since 2021 and am comfortable with Python, C, and Lua (for game engine scripting). In my free time, I enjoy reading and playing football. Below, you'll find a list of projects I have completed and certifications I have received.

Projects and Certifications

EDA of Spotify and YouTube Data

Conducted Exploratory Data Analysis of the top songs of a year on YouTube and Spotify using Python and data-analysis and visualization libraries (Pandas and Matplotlib). This included:


  1. Cleaning and normalizing the data
  2. Extracting useful features and removing non-usable features
  3. Plotting trends (e.g., views vs. song duration, views vs. likes ratio)
  4. Extracting meaningful relationships from the data
  5. Identified trends in user engagement
  6. Analyzed differences in audience preferences between platforms etc.


The data was sourced from here.

Machine Learning Specialization Course

Completed Andrew Ng's Machine Learning Specialization Course on Coursera and learnt the math and the fundamentals behind the following concepts:


  1. Main differences between Supervised and Unsupervised learning algorithms
  2. Linear and Polynomial Regression and associated loss function
  3. Logistic Regression and associated loss function
  4. Feed Forward Neural Networks and implementing them in Python using Tensorflow
  5. Decision Trees and related algorithms (XGBoost, Random Forest)
  6. Anomaly Detection using Normal Distributions
  7. K-Means Clustering
  8. Basics of Recommender Systems

Building my Own Neural Network

Created my own Feedforward Neural Network from scratch in Python, with only the help of mathematical libraries like Numpy. It has the following features:


  1. Variable number of layers
  2. Variable number of neurons in layers
  3. Multiple activation functions available for use - Linear, Sigmoid, ReLU, Leaky ReLU
  4. Multiple loss functions available for use
  5. Backpropagation (from scratch)
  6. Logging and graphing the learning rate

This helped strengthen my knowledge of the math behind Feed Forward Neural Networks, especially Backpropagation.

Academics

  • Degree: Bachelor of Science in Physics
  • Institution: Shiv Nadar University
  • Year: 2022–2026
  • Achievements: Dean's List, Research Projects in Machine Learning

Contact

If you'd like to get in touch, you can reach me through: