I’m Ahmed Farah, a Data & Software Architect focused on building end-to-end data and ML/NLP systems that turn raw inputs into production-ready products. Over the years, I’ve designed ingestion and transformation pipelines in Python, fine-tuned speech and language models for real corpora, and built semantic search and recommendation experiences on top of PostgreSQL (pgvector) and Elasticsearch. I also enjoy engineering the whole platform around the models—APIs, databases, scalable microservices, and self-hosted inference—while improving data quality with automated validation and dataset versioning. Currently, I’m developing a Tunisian dialect speech platform, creating training-ready audio datasets, fine-tuning voice models, and deploying quality gates that ensure reliable outcomes for downstream applications.…I’m Ahmed Farah, a Data & Software Architect focused on building end-to-end data and ML/NLP systems that turn raw inputs into production-ready products. Over the years, I’ve designed ingestion and transformation pipelines in Python, fine-tuned speech and language models for real corpora, and built semantic search and recommendation experiences on top of PostgreSQL (pgvector) and Elasticsearch. I also enjoy engineering the whole platform around the models—APIs, databases, scalable microservices, and self-hosted inference—while improving data quality with automated validation and dataset versioning. Currently, I’m developing a Tunisian dialect speech platform, creating training-ready audio datasets, fine-tuning voice models, and deploying quality gates that ensure reliable outcomes for downstream applications.WWWW…