Back
Personal Projects
SteamVault
Modular backend system to track Steam activity and generate daily snapshots, summaries, and insights for historical analysis beyond Steam's default playtime view.
Implemented a production-ready FastAPI backend with Supabase PostgreSQL and a custom caching layer, reducing redundant API calls and improving response performance.
Built analytics modules including 14-day trend calculations, activity heatmaps, multi-game comparisons, and streak detection, providing transparent metrics for user activity patterns.
Secure admin-protected routes and scheduled cron workflows (Google Cloud Scheduler) for automated fetch cycles, daily summaries, and uptime monitoring.
Designed a demo mode with a separate database to safely showcase analytics without exposing real Steam data.
Python
FastAPI
PostgreSQL
Supabase
Cron Jobs
Data Analytics
GitHub
Live Demo
Safecord
AI-assisted Discord moderation bot that helps identify potentially unsafe or predatory behavior through natural language processing (Zero-Shot Learning).
Built a FastAPI backend for AI analysis and data management, paired with a Discord.py-based command interface for real-time moderation.
Implements slash commands for flagging users, viewing watchlists, clearing logs, and running live message analysis using NLP.
All logs are securely stored in a local SQLite database, with a demo Jupyter notebook for exploratory data analysis and visualization.
Developed as a research prototype to explore proactive approaches to online safety and moderator empowerment.
Python
Fast API
Zero-Shot Learning (NLP)
SQLite
Discord.py
Data Analytics
GitHub
Breast Cancer Detection Model Comparison
Conducted a comparative study between Decision Tree (J48) and Neural Network (Multilayer Perceptron) models using the Weka Breast Cancer dataset.
Cleaned and preprocessed the dataset, handled missing values, and normalized features to enhance prediction quality.
Performed cross-validation and evaluated key metrics such as accuracy, precision, recall, and F1-score to assess performance trade-offs.
Used statistical tests to verify significance of differences and produced a professional analytical report with results visualizations.
Python
Weka
Machine Learning
Data Analytics
GitHub