m
muhammadmuaaz92

Muaaz Tahir

@muhammadmuaaz92
Pakistan
Anglais, Ourdou
Certaines informations sont présentées en anglais.
À propos de moi
Final-year Software Engineering student in Pakistan, working with Python, pandas, and scikit-learn on real ML projects. I recently completed an ML internship cleaning and processing production-scale data (~79M rows across 18 files) to build a scoring model, handling messy inputs, missing values, and data type issues. Currently in the NIAI-NAVTTC AI training cohort at NETSOL Technologies. On Fiverr, I clean, structure, and prepare datasets ready for analysis or modeling, and build simple classification models when needed. Send your file or project details and I'll take it from there. ... Plus d’infos

Compétences

m
muhammadmuaaz92
Muaaz Tahir
hors ligne • 
Temps de réponse moyen de 1 heure

Voir mes services

Nettoyage des données
I will clean and preprocess your messy dataset using python and pandas

Expérience professionnelle

Flyrank_AI

INTERNSHIP

Flyrank AI • Temps plein

May 2026 - Aug 20263 mos

Built a content opportunity scoring system as part of a remote ML internship (Lane 2: Refresh/Content Opportunity Scoring). Worked with 18 monthly data files (~79M rows) via a Hugging Face-hosted warehouse, using DuckDB for large-scale querying. Designed a proxy label and evaluation metric (precision@k), built a baseline rule-based scorer, then trained a logistic regression model with a time-aware train/test split. Identified and fixed a data quality bug where a key feature was silently stored as an invalid type, and used a fairer evaluation subset to properly measure real model skill (AUC 0.782) after catching a flaw in the baseline's inflated precision.