NastaRead
OCR pipeline for degraded Urdu Nastaleeq text with restoration, transcription, and API deployment. Parallelized inference using Hadoop MapReduce in a fully containerized Docker environment, enabling horizontal scaling for large document batches.