Skip to content
View loh-0's full-sized avatar

Block or report loh-0

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
loh-0/README.md

πŸ‘‹ Hi, I'm Lohith Karthesan!

πŸŽ“ I'm a Computer Science student at Monash University, specialising in Data Science.
πŸ“Š Passionate about turning data into insights and driving social impact through technology.
πŸ’‘ I love solving real-world problems and building tools that make a difference.


Tools & Technologies

Programming

  • Python
  • R
  • SQL
  • Java

Featured Projects - Click to Try it out!

  • Built a fully interactive Streamlit web app to predict loan approvals using a Random Forest model.
  • Integrated SHAP explainability tools to visualise individual prediction reasons and feature impacts.
  • Designed a What-If simulator to explore how improving CIBIL scores or reducing loan amounts could improve approval chances.
  • Highlighted data limitations: model mainly relied on CIBIL score and loan term due to limited feature influence.
  • Developed supervised learning models on 250,000+ transactions to identify fraud.
  • Used SMOTE to balance the data and improve recall of fraudulent classes.
  • Evaluated with confusion matrix, increased true positive rate by 11%.
  • Developed betting strategy optimisation using Markov Chains and simulated annealing, improving win rates from 45% to 96% by modeling stochastic transitions between financial states
  • Conducted Monte Carlo simulations with 10,000+ iterations to validate analytical results and analyse probability distributions for different starting conditions and strategies
  • Applied matrix exponentiation and absorption probability calculations to determine expected outcomes and time-to-absorption across different balance states in a discrete stochastic system
  • Built transition matrices and implemented optimisation algorithms to solve strategy selection problems, demonstrating quantifiable performance improvements through statistical analysis
  • Built a full-stack paper-trading simulator using React, Node.js, and MongoDB Atlas.
  • Simulates real-time NASDAQ trading using Finnhub API for live stock prices.
  • Features include portfolio tracking, buy/sell flow, profit/loss calculations, and position management.
  • Backend hosted on Render, frontend on Vercel, with configs handled via environment variables.
  • Designed a professional UI with responsive layout, animated notifications, and smart error handling.

πŸ“« Get in touch!


⭐ Thanks for stopping by my GitHub!

Pinned Loading

  1. Credit-Card-Fraud-Detection Credit-Card-Fraud-Detection Public

    Jupyter Notebook

  2. Markov-Chain-Based-Betting-and-Strategy Markov-Chain-Based-Betting-and-Strategy Public

    Jupyter Notebook

  3. Loan-Approval-Prediction Loan-Approval-Prediction Public

    Jupyter Notebook

  4. Trading-Sim Trading-Sim Public

    JavaScript

  5. sleep-application sleep-application Public

    Sleep Tracking Project

    Python