← Back to Home
Projects

Building and automating, one experiment at a time

I love building and automating solutions that blend software, data science, and industrial processes — from a full desktop analytics app to focused one-off models and dashboards.

Featured Project
Desktop App · Windows

theSA — The Structured Analytics

My biggest build so far — a full desktop app for the whole life of a dataset, written to remove the friction between cleaning, modeling and visualizing data across separate tools.

  • Data Cleaning — fix nulls, duplicates, and outliers, then send the cleaned set straight to any other module.
  • Predictions & Anomalies — train and compare classification, regression, and outlier-detection models side by side.
  • Time Series — pattern detection, change-point detection, and forecasting with confidence intervals.
  • Sessions — save the entire working state, every module and dataset, and resume later.
Explore theSA →
SA
Data Cleaning Machine Learning Time Series Visualization
Data & AI

Models, dashboards & a chatbot

Local PDF Chatbot

GitHub Repo → 📎 View PDF
  • Offline RAG pipeline that loads and indexes every PDF in a local folder — text and images both.
  • Uses BLIP-2 to caption images and pulls surrounding text for context before embedding.
  • Stores chunks in a FAISS vector database, retrieves the top matches, then reranks with a BGE reranker.
  • Sends only the top reranked chunks to a local LLM and answers with source PDF name and page number.
  • Skips already-processed files and runs fully offline, no cloud APIs required.

Credit Card Financial Dashboard

GitHub Repo → 📎 View PDF
  • Interactive dashboard built on transaction and customer data from a SQL database.
  • Streamlined data processing to monitor key performance metrics and trends in real time.
  • Shared actionable insights with stakeholders to support decision-making.

Sales Report Dashboard

GitHub Repo → 📎 View PDF
  • Time-series analysis for accurate sales forecasting and an interactive reporting dashboard.

Email Spam Detection ML Model

GitHub Repo →
  • Python ML model that classifies whether an email is spam or not.

Heart Disease Prediction Model

GitHub Repo →
  • A classification model that predicts whether a person has heart disease.
More

Concepts, this site, and the rest

Machine Learning Concepts

GitHub Repo →
  • A running collection of different models and concepts in machine learning.

My Website Project

GitHub Repo →
  • The source for this very site — see how it was built in VS Code.

Explore My Other Works

GitHub Repo →
  • The full repository list, for everything not featured here.