Remin Franklin Eliyas

AI Engineer specialising in MTC extraction and OCR model evaluation. Responsible for testing multiple document intelligence solutions, benchmarking performance across accuracy, latency, and robustness metrics, and delivering structured comparative analyses to support model selection and optimisation.

Remin focuses on the practical challenge of making AI work reliably on real manufacturing documents — mill test certificates, material data sheets, and compliance records that vary enormously in format, quality, and completeness. His work directly supports GoSmarter’s MillCert Reader product, which automates the extraction of chemical and mechanical property data from steel certificates.

His evaluation process is rigorous: testing multiple OCR and document intelligence models against the same datasets, measuring accuracy on structured fields, tolerance for poor scan quality, and the consistency of outputs across certificate types from different mills and standards bodies. The goal is not just finding the model that scores highest in a benchmark, but finding the one that production teams can actually trust.

Remin holds a Masters in Data Science and brings a methodical, evidence-based approach to model selection. He is part of the core engineering team at GoSmarter and works closely with the product and customer success functions to ensure AI performance translates into real-world reliability for manufacturers.