Ricardo Fitas
Contact: rfitas99@gmail.com
I work as a Physics Expert at Mercor, preparing training data for large language models by solving physics problems. My background includes an MSc in Mechanical Engineering, credentials in Statistics and Data Science from Massachusetts Institute of Technology (MIT), research experience at TU Darmstadt, the Technical University of Munich (TUM), RWTH Aachen University, the University of Porto, and a European research grant through the University of Aveiro. My research combines computational mechanics, data science, machine learning, numerical analysis, and optimization. My publication record comprises more than 20 research papers, collectively cited over 350 times. Teaching and mentoring include supervision of more than eight BSc and MSc theses and hundreds of hours of private tutoring for students at highly ranked universities. Scientific service includes over 100 peer reviews for journals published by Elsevier and Springer Nature, alongside membership of technical program committees for IEEE conferences. Further experience spans research project management and applied collaborations with academic and industrial partners.
Research Interests
Self-Discovery Lab
Autonomous scientific research that selects questions, develops hypotheses, executes computational experiments and learns from reproducible evidence in mechanics, materials and scientific machine learning.
Physics-informed Machine Learning
Neural models that respect differential equations, constitutive behavior and thermodynamic constraints, with particular interest in solid mechanics and inverse problems.
Computational Solid Mechanics
Finite element modeling of deformation, contact, nonlinear response and material behavior, supported by experimental evidence.
Constitutive Modeling & Multiscale Transfer
Understanding how material realizations and internal states transfer between scales, geometries and loading paths.
Metamaterials & Inverse Design
Exploring how architecture controls mechanical response, and designing structures to meet specified stiffness, deformation and performance targets.
Digital Twins & System Identification
Connecting simulations to measurements through calibration, parameter identification and models of physical experiments.
Stochastic Systems & Uncertainty
Quantifying uncertainty, testing robustness and using stochastic modeling to understand the reliability of engineering predictions and decisions.
Engineering Data & Graph-based Models
Reproducible datasets and graph representations of physical structures, including geometry tracking and analysis of corrugated-board experiments.
Structural, Multi-objective & Learning-based Optimization
Structural and topology design, particle swarm methods, genetic hybrids and learned optimization strategies for exploring Pareto trade-offs in engineering design.
Materials, Manufacturing & Sustainable Packaging
Linking material properties, manufacturing quality and structural design to improve resource-efficient components and corrugated packaging.
