Skip to content

Open to full stack, cloud and AppSec roles

Taimur QaziFull stack engineer who builds secure web platforms on AWS.

3+ years shipping production apps for trading, construction, travel and AI products, from React and Next.js front ends to Node services, Terraform and CI/CD. Now finishing an M.S. in Cybersecurity at the University of Delaware.

View projectsRésuméGet in touch
Years in production
3+
Projects, 7 live
18
M.S. Cybersecurity GPA
3.94
CompTIA certified
Sec+
ProjectsExperienceEducationToolboxTXTabout me.txtPDFresume.pdfcontact meterminal
Bank Fraud Detection

Bank Transaction Fraud Detection, ML team project | Taimur Qazi

Bank Transaction Fraud Detection

Machine Learning team project, University of Delaware (Fall 2025)

2,512bank transactions
7models compared
5engineered fraud signals

A Python model that flags suspicious bank transactions and gives each one a fraud probability, built with a classmate for the Machine Learning course.

  • Engineered fraud signals from raw transaction data: high amounts, repeated login attempts, rapid repeat transactions, unusually long transaction times and channel patterns.
  • Labeled the data with a rule that flags any transaction tripping 3 or more signals, giving a 10% fraud rate, similar to real data.
  • Trained and compared 7 models (logistic regression, linear and multiple linear regression, a decision tree, SVMs with linear and RBF kernels, and k nearest neighbors), scored on accuracy, precision, recall, F1 and ROC AUC.
  • Recommended logistic regression for its interpretability and fast real time scoring.
Pythonpandasscikit-learnMatplotlibJupyterFeature engineering
Academic project, no public buildAsk for a walkthrough