I worked as a Data Engineering Intern at Shiftbricks, a start-up where I developed a data pipeline following the Medallion architecture (bronze, silver, and gold layers) to process Arabic legal documents. The pipeline extracted and structured metadata and content from PDFs using a combination of OCR, LLM models, and other advanced techniques, converting the data into a JSON schema for storage in MongoDB. The structured content was then embedded for use in an AI assistant chatbot.This entire process was automated with Airflow and included a human validation step through a FARM stack UI, where experts validated the document structure before embedding. Docker was used to fully containerize the pipeline, ensuring a smooth and scalable deployment.