Summary
Marcos De Souza is an AI/ML researcher and software engineer with over a decade of experience building scalable, production-ready solutions and teaching at university level. Currently an invited assistant professor and researcher at the University of Coimbra and CISUC, he specializes in unsupervised learning, transformer-based autoencoders and industrial computer vision—recently applying ViTAE and YOLOv11 to automate cork-inspection with a strong focus on low-latency inference, interpretability and MLOps deployment. His background spans R&D leadership in industry (Inatel, CESAR) and applied product work from mobile/embedded vision to cloud-based ML platforms, giving him a rare mix of academic rigor and pragmatic engineering. He holds a PhD in Machine Learning and an MBA-informed appreciation for business constraints, which helps him translate complex models into deployable, cost-effective systems. An understated strength is his track record of optimizing for real-world constraints (small labeled datasets, edge inference) rather than chasing raw benchmark performance.
10 years of coding experience
12 years of employment as a software developer
Software Engineer Specialization Software Engineering, Software Engineer Specialization Software Engineering at Universidade de Pernambuco
Doctor of Philosophy - PhD Machine Learning, Doctor of Philosophy - PhD Machine Learning at Centro de Informática UFPE
MBA Business Administration and Management General, MBA Business Administration and Management General at UNIFG
Bachelor's degree Information systems, Bachelor's degree Information systems at Faculdade Joaquim Nabuco
Portuguese, English