Innovative Data Scientist with Computational Mechanics Experience: Journey from Academia to FinTech

Photo courtesy of Minghao Liu

The opinions expressed by Digital Journal contributors are their own.

Minghao Liu’s career is a testament to the power of computational science in solving complex real-world problems. Beginning with his pioneering research in computational mechanics, Minghao has consistently demonstrated a passion for pushing the boundaries of understanding of dynamical systems, especially under extreme conditions where traditional experiments fail. His experience in developing advanced simulation and computational models not only filled critical gaps in scientific knowledge, but also laid the foundation for his smooth transition into the world of FinTech.

His doctoral dissertation, which resulted in many influential publications, reflects a desire to expand the boundaries of understanding through modeling.

Peridynamic modeling of brittle ice crushed by a vertical structure

In this study, Minghao applied the theory of peridynamics to model the fracture mechanics of brittle ice subjected to impact forces, conditions that are difficult to reproduce in physical experiments. The work provided important information about crack propagation patterns in extreme cold conditions, which has helped improve the design of structures exposed to harsh arctic conditions.

Coarse-grained molecular modeling of the microphase structure of a polyurea elastomer

In this paper, Minghao developed a structure corresponding to the coarse-grain model of polyurea elastomers. The model implicitly represented hydrogen atoms, simplifying the complexity of molecular modeling. Trained using iterative Boltzmann inversion and a new distance-dependent scaling function, the model significantly reduced iteration time. The simulation captured the microphase separation of polyurea, revealing a rigid domain spacing of five nm, consistent with X-ray scattering data from similar elastomers. By analyzing two different model systems, the study found that multi-block systems form large, interconnected rigid domains, while two-block systems create smaller, ribbon domains. The study also showed that the soft segments formed bridge-like and loop-like structures, which contributed to a better understanding of the microstructure of the material under different conditions.

Coarse-grained molecular modeling of the role of cure rate on polyurea structure and strength.

This study examined how the curing process affects the mechanical strength of polyurea, a material that plays a key role in industries requiring high-performance materials such as defense and automotive. Modeling these effects in silico allowed us to understand in more detail how material properties change under extreme production conditions.

The use of scientific computing was central to all of Minghao’s research efforts. Using computer simulations, Minghao was able to simulate complex physical systems at multiple scales, from fracture mechanics in ice to molecular dynamics in polyurea. The power of high-performance computing has made it possible to study system behavior that is difficult or impossible to observe in real experiments. These computational methods, such as peridynamics, coarse-grained modeling, and lattice Boltzmann, have not only reduced the need for physical prototypes, but also accelerated the discovery process by providing greater insight into the properties of materials under extreme conditions.

This background in scientific computing easily translated into Minghao’s current role as a data scientist in the financial technology industry. In a field where large-scale data analytics and machine learning are driving innovation, Minghao is applying similar computational methodologies to solve financial problems. Just as modeling was used to predict the behavior of materials, predictive models and algorithms are now used to optimize financial products, analyze market trends, and improve the customer experience.

Specifically, in his data science role, Minghao specializes in credit risk modeling and recommendation systems. The skills developed through his academic research—modeling dynamic systems, analyzing complex data, and making forecasts based on those models—are directly applicable in a financial context. For example, modeling credit risk requires understanding the patterns of financial behavior of trade data under different conditions, much as Minghao once modeled the material properties of molecular structures under extreme stress. Similarly, building recommender systems involves predictive modeling and pattern recognition, skills he honed working with molecular and materials modeling. In both cases, Minghao offers a unique approach that combines scientific rigor with innovative financial solutions.

Combining scientific rigor with innovative financial solutions, Minghao Liu demonstrates how advanced computational methodologies can transcend disciplines, driving both technological and business innovation. His journey from academia to FinTech highlights the transformative power of data science and computational knowledge in solving the toughest problems across industries.