fruad-detection

An end-to-end Machine Learning project for credit card fraud detection, covering data analysis, preprocessing, model training, evaluation, and deployment.

ML-powered fake job detector — Linear SVM + DistilBERT ONNX, 10-signal URL scorer, runtime model switching.

📝 Logistic regression fraud classifier on 284K+ transactions — 87% accuracy, 94% AUC-ROC

A free, bilingual financial literacy and decision-support platform helping users understand loans, discover government schemes, learn financial concepts, and identify potential fraud risks.

A SIEM-based Fintech Threat Monitoring project using Splunk Enterprise to investigate suspicious transactions, phishing indicators, transaction structuring, malicious IP activity and threat hunting through interactive dashboards and SPL queries

End to end pipeline detecting product fraud and IP infringement in marketplace listings. Built in public in 21 days

KYCShield is a cybersecurity research project focused on protecting electronic Know Your Customer (eKYC) systems against deepfakes, presentation attacks, replay attacks, and synthetic speech. It combines video forensics, audio anti-spoofing, liveness verification, and multimodal risk analysis to support more secure digital identity verification.

End-to-end ML platform for credit risk & fraud detection: 9 models (incl. TabNet), SHAP/LIME explainability, MLflow tracking, FastAPI serving, Docker deployment. 96% test coverage.

AI-powered scam detection platform built with Reflex, Python, Supabase, Groq AI, Docker, and Railway. Detect fraudulent messages, URLs, QR codes, and screenshots in real time.

Flags suspicious transactions for AML review using rules, graph patterns, and an LLM for the final call.