Payroll Automation
Python tool that pulls timesheets via API, applies pay rules (OT/holidays/rates), and exports payroll-ready reports. Added a one-click GUI so anyone can run it.
Read case study →Everything Computer Science
I love learning, creating, and building tools that help people!
Python + simple UIs for real business workflows: payroll, data cleanup, form automation, and review pipelines.
About
I’m Josie, a software developer (UMass Amherst, CS) focused on building clean, reliable internal automations. I like turning messy, manual processes into one-click runs with clear logs and easy handoff. Right now, I’m deepening my understanding of database fundamentals and working to master web analytics and scraping.
Projects
Database Cleaner
Audited & mapped Salesforce data, removed redundancy, and aligned structures to support Conga form automation. Outcome: more reliable document generation and simpler future data entry.
Automating Quality Control
Scraped public reviews and internal docs, normalized into review categories for a QC queue. Enabled faster triage and trend tracking.
Spanish Language Learning Chatbot
Final project for my AI course at UMass Amherst, created with teammates Shachi Benara and Devansh Jain. We built a chatbot that helped users practice beginner-level Spanish by presenting sentences in English or Spanish and prompting responses in the other language. The system scored answers out of 10 and tracked progress across lessons to measure improvement over time.
Tech: Python, Pandas
Payroll Automation — Case Study (Safe Demo)
Problem
Payroll prep involved manual copy/paste from time tracking, rate lookups, and special-case rules. It was slow and error-prone.
Solution
- Pull timesheets from a time-tracking API.
- Apply rules for OT/holidays/differential rates.
- Export payroll-ready CSV/Excel with audit log.
- One-click GUI for non-technical users.
Results
- ~80–90% time saved per run.
- Fewer corrections; repeatable & auditable.
Tech
Python, Flask, Pandas, REST APIs, CSV/Excel.
Redacted Screenshot
Architecture Diagram
This page uses synthetic data and redacted visuals. No client names, no real employees, no sensitive fields.