Projects + Explorations

Exploration & Creativity

This is where I'll be sharing some of my latest experiments, creative projects, and cool finds!

How I Built My Resume

If you've reviewed my resume or are reviewing it now, you'll know I've always hated fighting with templates. So I decided to build my own. In this article, I walk through how I created a clean, LaTeX-style, ATS-friendly resume using React, Cursor, and AI tools.

Read: How I Built My Resume

How I Built a 0-to-1 Web App with Cursor

I wanted to see how fast I could go from idea to working product, so I used Cursor to build a web app from scratch in just a couple of hours.

Read: Building a 0to1 Web App with Cursor

Previous Projects

Here' a collection of projects I worked on during grad school, organized into three sections: Data Science, Data Visualization, and Case Studies. It's a mix of hands-on work and creative problem-solving from that time!

Data Science Projects

E-Commerce Customer Segmentation & Behavior Analysis

  • Predicted product recommendations depending on past 259k customer transactions; used K-means Clustering to aggregate frequently bought items together, which lead to a strategy for promoting market basket sales improving revenue on retail
  • Performed RFM Analysis, to segment customers based on value from past transactions and Trend analysis to find purchase patterns over time
PythonMachine LearningK-meansRFM Analysis

ARTICULATE : Article Summarizer

  • Web-Scraped data using Selenium to fetch articles from NY Times, summarized article lengths using techniques like TextRank, Similarity matrix
  • Classified the comments on articles to positive/negative and truthful/deceptive using Sentimental analysis and Text analysis
  • Predicted NYT's best pick comment using Logistic regression with 72% ROC; built an automatic comment generator using Tensorflow and LSTM
NLPMachine LearningPythonTensorflow

Moderation System for Hate Speech Detection

  • Designed an interface that identifies abusive content which alerts users, and provides a timer to rethink before publishing hate content using BERT
  • Performed data cleaning using Lemmatization, Stemming and did feature extraction with techniques like TF-IDF vectors, Lime Text Explainer
  • Used oversampling- SMOTE and RandomOversampling on the imbalance data to achieve the accuracy of 95%
BERTNLPPythonMachine Learning

A/B test on Mobile Game Players

  • Conducted A/B test on mobile game data with ~90k players to determine the intensity of obstacles in-game level to increase player retention
  • Implemented the hypothesis testing using statistical techniques like Chi-Square test, degree of freedom, p-value
Statistical AnalysisPythonA/B TestingChi-Square

HR Employee Management

  • Applied HR Analytics using ML on IBM's dataset, created a strategic plan for employee churn model & retention rate
  • Reduced attrition rate by determining KPI's & flagging employees at risk; Logistic regression, SVM with 82% accuracy
Machine LearningPythonLogistic RegressionSVM

LinkedIn Messages & Connection Analysis

  • Performed analysis on my current LinkedIn Messages and connections for detecting language & to determine the tone of msgs I've been getting.
Data AnalysisPythonNLP