HEY, I'M AJAYVEER
A DUAL MAJOR COMPUTATIONAL MATHEMATICS & STATISTICS STUDENT AT THE UNIVERSITY OF WATERLOO. DRIVEN BY CURIOSITY AND A PASSION FOR PROBLEM-SOLVING.
MY PORTFOLIO
A SHOWCASE OF MY RECENT WORK AND CREATIVE SOLUTIONS
ZocraticMMA
ZocraticMMA is an app I have created with my friend to give UFC fans a smarter, data-driven way to explore fighters, compare matchups, and predict outcomes. With 20+ monthly users, I developed custom web scrapers to collect fight data, stored everything in a structured SQL database, and created APIs so the app can access and deliver that data in real time.
KEY FEATURES
TECH STACK
LIVE APPLICATION
ZOCRATICMMA.COM
Canadian Unemployment Forecasting (Macro ML)
I explored whether Canadian unemployment can be predicted using macroeconomic indicators and how interest-rate decisions feed through into the labour market. Using monthly data from 2009–2024, I built and evaluated a forecasting pipeline around several baseline, time-series, and machine-learning models. The project walks from data cleaning and feature engineering through model selection, diagnostic checks, and scenario design.
KEY FEATURES
TECH STACK
ACTUAL VS PREDICTED
IMPACT & RESULTS
Shows that simple, interpretable models (especially Persistence and Ridge) can match or beat more complex ML at a 1‑month horizon for Canadian unemployment. The scenario module turns interest‑rate paths into plausible unemployment trajectories, making the work useful as a 'what if' policy tool and a demonstration of end‑to‑end econometric modeling.
Movie Recommendation Tool
I built a personalised, interactive movie recommendation tool to help friends and casual movie watchers save time and avoid endless scrolling. I noticed how often people, especially students and busy professionals, struggled to pick a movie quickly, so I designed a solution using Python tailored recommendations based on their preferences.
KEY FEATURES
TECH STACK
TECHNICAL CHALLENGES
Processing user preferences like genre, era, and runtime to provide accurate recommendations while maintaining a simple and intuitive user interface.
IMPACT & RESULTS
The tool narrows decision-making by processing the user's preferences like genre, era, and runtime, then returning recommendations based on them. I showed my software to my friends, and they reported saving 15+ minutes when using the app.