Marco Molinaro

Marco Molinaro is a Professor in the Computer Science Department at PUC-Rio. He previously held Assistant Professor positions at TU Delft (EWI), Georgia Tech (ISyE, ACO), and was a PhD student in the ACO program at Carnegie Mellon University.

Education:

• PhD in Algorithms, Combinatorics, and Optimization (ACO), Carnegie Mellon University

Research Interests: Marco's work lies at the intersection of online algorithms, convex optimization, machine learning, and operations research. His recent publications analyze:

• Online scheduling and VM allocation in data centers
• Complexity of branch-and-bound trees
• Information-theoretic bounds for optimization problems
• Explainable decision trees and their theoretical limits
• Stochastic and adversarial bandit algorithms
• Curvature properties in optimization sets

His research often bridges abstract mathematical frameworks (e.g., Lipschitz functions, supermodular norms) with practical applications in cloud computing, machine learning, and game theory.

Publication Trends: Marco's recent work focuses on online convex optimization and its extensions to data center resource allocation (Kamino: latency-driven VM scheduling), dynamic bin packing, and information complexity in mixed-integer models. He frequently collaborates with researchers at institutions like Carnegie Mellon University and TU Delft, with recurring co-authors Santanu Dey, Amitabh Basu, and Thomas Kesselheim.

Teaching: Marco teaches advanced courses in algorithm design and analysis at PUC-Rio, including:

• Analise de Algoritmos (2020–2025)
• Projeto e Analise de Algoritmos (postgraduate, 2020–2025)
• Algoritmos e Incerteza (postgraduate, 2020)
• Estruturas Discretas (2021)

His curriculum emphasizes online learning, random-order models, and PAC learning, with connections to game theory and practical implementations.