Hi, I’m Siddhi.

I build practical data and machine learning solutions rooted in research. Explore my projects below, and feel free to connect!

A document retrieval system that uses a hybrid model combining TF-IDF-based Vector Space Modeling (VSM) with Single-Link Hierarchical Clustering to semantically group and rank documents based on query similarity. It features a minimal Streamlit interface for uploading, searching, and visualizing clustered results, with MongoDB used for document metadata management. The methods and approach implemented in this project are detailed in a research paper authored by us and published in IEEE. Research Paper (IEEE)

Demo:

Demo:

SentimentFusions is a full-stack product review analyzer that scrapes real-world reviews, processes them using machine learning, and displays a sentiment dashboard with keyword insights and rating trends — all on a single interface.
Built for fast, reliable customer opinion mining across any product or category.

An end-to-end machine learning project to detect fraudulent financial transactions using a stacking ensemble model (Random Forest + XGBoost). It includes preprocessing, modeling, evaluation, prediction, and a user-friendly Streamlit web app.

Demo:

An exploratory data analysis (EDA) project on restaurant data from Zomato in Karnataka, India. It includes data cleaning, feature engineering, and an interactive Power BI dashboard to uncover insights into restaurant trends, ratings, and costs.

Demo: