Youssef Zerbouh

Web Scraping Pipeline Designed and automated a scalable BrightData proxy pipeline to scrape, parse, clean, and deliver structured datasets within a 48-hour turnaround. Implemented data extraction workflows, parsing logic, and ETL automation to ensure reliable and high-quality data delivery for enterprise use cases.

Stack: Python, BrightData, Web Scraping, ETL Pipelines, Data Automation

LLM Training Data Enterprise Built and structured agent interaction datasets with validation and preprocessing layers to support enterprise-scale LLM fine-tuning workflows. Developed AI data engineering pipelines focused on dataset quality, consistency, and scalable processing for machine learning applications.

IoT Vehicle Data Pipeline — Built a real-time IoT ingestion and ETL pipeline for vehicle sensor streams, enabling scalable telemetry processing and data enrichment.

Stack: Python, Pandas, ETL, IoT Edge

Containerized Data & ML Platform — Containerized and deployed the full ingestion, ETL, and ML serving stack using Docker and Kubernetes on DigitalOcean with a Flask REST API.

Stack: Docker, Kubernetes, Flask, DigitalOcean

Predictive Maintenance Model — Developed and deployed an Attention-based RNN achieving 86% failure prediction accuracy, optimized with Bayesian hyperparameter tuning.

Stack: PyTorch, Python, Deep Learning, Bayesian Optimization