Projects and case studies
Six systems I built and shipped: a live RAG assistant, an OCR pipeline measured against Tesseract, a fail-closed monitor from my internship, and three smaller builds. Each page shows what broke, what was measured, and what still does not work. For a source-linked index of project measurements, see the evidence page.
AI & RAG
UET GPT
AI assistant for university students
An unofficial assistant for UET Taxila students and applicants. It retrieves the university's published pages for questions about admissions, merit, fees and programs, with citations, refusal instructions for empty retrieval and hedging for low scores. Not affiliated with UET Taxila.
Document Intelligence
B.L.A.S.T.
Reads text out of documents (OCR)
Extract text from a PDF, PowerPoint or image into Markdown or Word. Across versions adding layout handling and changing the OCR engine, its character error rate on a 14-page test set fell from 0.499 to 0.108.
Automation & Monitoring
Marketplace Monitoring & Alerting
Watches freelance marketplaces and emails new jobs
Watches freelance marketplaces for new postings and emails a digest after each successful scan. When a dependency fails it stops and says so instead of reporting an empty result. Built during my internship at Paramount Intelligence.
Menu Data Pipeline
Pulls 17 restaurant menus into one live spreadsheet
Concurrently pulls menu data from 17 restaurant sites, validates and cleans it with Pydantic and pandas, and pushes a unified dataset straight into a live Google Sheet.
Engineering Experiments
Nightshade
Obfuscates Java code against AI training scrapers
A tool that rewrites your code so it still compiles and runs correctly for humans and compilers, intended to make scraped copies less useful for AI training.
Quantum Singularity
WebGL particle art you control with your hands and voice
A real-time WebGL2 particle simulation (up to 110,000 GPU particles) driven by hand gestures and live audio, running entirely in the browser from a single static HTML file.
