Fraud Detection Machine Learning algorithms on the Credit Card Fraud Dataset. Balancing classes with synthetic samples (SMOLE)

Graph-based analytics and HGT models to identify suspicious transaction networks (fraufulent transaction in fintech and digital wallets)

Embedding transaction graphs to classify fraudulent Ethereum wallet addresses.

This is a comprehensive fraud detection system built using machine learning to identify fraudulent financial transactions in real-time. The project implements a Random Forest classifier trained on the PaySim dataset, handling highly imbalanced data using SMOTE (Synthetic Minority Over-sampling Technique) and engineered features with 99% accuracy.

Supervised ML pipeline for detecting suspicious transactions using a synthetic AML dataset. Includes EDA, feature engineering, SMOTE, and model evaluation.

his repository contains a Streamlit dashboard for visualizing fraud detection insights based on credit card transaction data.

An end-to-end AI fraud detection platform using an XGBoost model, LangGraph agent workflow, FastAPI, and a full MLOps CI/CD pipeline with Docker and GitHub Actions.

Kaggle Fraud Detection challenge hosted by IEEE

Fraud detection with tensorflow.

Real-Time Log Processing API built with FastAPI, Celery, Redis, and MongoDB

A CTGAN variant with SimCLR-style NT-Xent contrastive loss for better synthetic credit-card fraud data. Evaluated on both data fidelity and utility via a XGBoost classifier.

Logistic Regression, Grid Search, and ROC-PR curve evaluation on fraud detection dataset

Synthetic Credit Card Transaction Generator used in the Sparkov program.

This is a fraud detection algorithm that is designed to detect fraudulent activity in online in-game markets using a combination of unsupervised learning and a rule based approach.

💳 Revolutionize payment systems with Zen7 Payment Agent, a DePA implementation that enables automated, secure transactions and innovative commerce solutions.

FraudShield leverages a semi-supervised learning framework to detect financial fraud. It combines supervised learning and anomaly detection to showcase effective fraud prevention in a Streamlit app, using the BAF Dataset for real-world scenario simulation.

Project Lumina is a collection of Fraud Detection algorithms using Graph Neural Networks.

Detecting fraudulent insurance claims using machine learning techniques like XGBoost and SMOTE on real-world data to enhance claim verification accuracy.

Portfolio Project: AI-driven financial transaction risk detection using automation workflows and real-time model scoring.

End-to-end automated modern ELT data pipeline using Python, PostgreSQL, Airflow, Kafka, GCS, BigQuery, and dbt — built to simulate a production-grade e-commerce environment. Completed as part of Purwadhika Data Engineering Bootcamp Final Project