Built a Random Forest classifier in Python to detect brain tumors from a medical imaging dataset. Identified and corrected data leakage issues in the original pipeline to ensure the model's evaluation results were reliable rather than artificially inflated.
→ github.com/ayesha-cs-93/brain-tumor-classificationBuilt an end-to-end ML pipeline with LightGBM and feature engineering (ratio features, ordinal/target encoding, BMI bucketing), validated with 5-fold stratified cross-validation. Reached a public leaderboard score of 0.95030 across 13 iterations and earned the "Code Uploader" badge.
→ github.com/ayesha-cs-93/student-health-risk-predictionA Supabase-backed authentication API built during the Backend AI Engineering track at FlyRank AI, implementing secure user auth flows with clean endpoint design.
→ github.com/ayesha-cs-93/auth-apiA full CRUD API built with FastAPI and SQLite, later containerized with a Dockerized PostgreSQL setup (BE-04) — built with a clean, staged Git commit workflow.
→ github.com/ayesha-cs-93/task-crud-apiA dashboard interface built with custom CSS grids, glassmorphism styling, and animations.
→ github.com/ayesha-cs-93/Podcast-Platform-UIPython, JavaScript, SQL
FastAPI, REST API design, SQLite, PostgreSQL, Docker, Supabase (Auth)
HTML5, CSS3 (animations & layouts), Bootstrap
Pandas, Scikit-learn, LightGBM, Feature Engineering, Cross-Validation, Exploratory Data Analysis
Git, GitHub, VS Code, Kaggle, Google Colab, Canva
• Kaggle Playground Series Competition — Top ~20%, public leaderboard score: 0.95030
• Earned Kaggle "Code Uploader" badge
• AI/ML Intern — FlyRank AI (Backend AI Engineering & AI Fluency tracks)
• Open Source Contributor — NSoC'26