At FlyRank, I built a RESTful task API in Express with CRUD operations and Swagger/OpenAPI documentation, then migrated its storage from in-memory to SQLite. I also re-implemented persistence in Python using FastAPI, asyncpg, PostgreSQL, and Docker.
At UneeQ Interns, I delivered three machine learning projects across recommendation systems, deep learning, and classification. My work included a hybrid movie recommender evaluated on 100K+ ratings and an MNIST classifier that achieved 98%+ accuracy.
At DEPI, I completed IBM's Data Science program and benchmarked forecasting models for Rossmann sales. XGBoost achieved 1.06% MAPE, with the pipeline tracked in MLflow.
My independent projects include Songbook, a web app for phonetic song guides across 19 scripts, and a semantic plagiarism checker using Sentence-BERT. I also built a real-time face anonymization pipeline and a jobs and recruitment automation dashboard.
