A. Postdoctoral Research Experience at Colorado State University (CSU), Oct. 2021 - Date

• Developing sparse optimization and ML tools to extract from biological data information (biomarkers) exclusive to and common to the nose and blood for targeted therapeutics.

• Analyzing high-dimensional biological datasets to identify biomarkers (genes) responsible for coronavirus (MERS-CoV) progression including up- and down-regulated genes.

• Developing and applying ML algorithms such as Support Vector Machines (SVM) to predict disease status as well as pathway discovery. 

• Integrating gene expression datasets to detect and analyze transcriptomic patterns of the host immune response to infection by monitoring changes in expression levels. 

• Developing data analysis ML toolbox for infectious disease mitigation - early warnings, surveillance, detection, spread, prevention, and control of zoonotic threats. 

• Developing tensor optimization and ML algorithms for multimodal analysis of big-omics noisy high-dimensional data sets with few samples, e.g., genomics data.

B. Doctoral Research Experience at Kent State University (KSU), Aug. 2018-21

• At KSU, my Ph.D. dissertation focused on Iterative Tensor Decomposition Based on Krylov Subspace-Type Methods with Applications to Image Processing. This work is concerned with developing more efficient and robust techniques for multidimensional data analysis. 

• Developing novel techniques for tensor data analysis is important because most real-world data are high-dimensional. Specifically, my Ph.D. dissertation uses multidimensional structures to store and encode information more efficiently. It developed optimization algorithms that process data multidimensionally to leverage data structures and used tensor decompositions to elicit feature extraction when data is naturally high-dimensional. 

• Tensor-based algorithms lead to more efficient and robust computational approaches to data reconstruction than traditional matrix-based algorithms because they preserve data structure by avoiding flattening (vectorization and matricization). Flattening often leads to loss of information and the destruction of spatial correlations and structural complexities inherent in data. 

• Applications, such as MRI image reconstructions and 3D reconstruction of CT scans, are some of the many broader impacts of my work.

C. Doctoral Experience as Wallace Givens Associate at Argonne National Laboratory (ANL), May - Aug. 2021

• Investigated stochastic (tensor) optimization frameworks for ptychographic imaging.

D. Master's Research Experience as a Visiting Intern at DNV GL, London, UK, June - Aug. 2015

• My M.Sc. thesis at the University of Bath focused on Modeling Flammable Effects – Fire! Throughout the duration of the project, I was a visiting intern at DNV GL, London between June - August 2015. 

• My M.Sc. thesis is titled: Radiation Modeling Near and Asymptotically Far Afield: A Case of a Pool Fire. This work focused on developing cost-effective approaches to compute the view factor, i.e., the configured amount of radiation that reaches an observer from the flame surface, for both near and asymptotically far-afield observers. 

• Pool fires are the most frequent of all possible types of fires around the world, and in every fire hazard, there are losses of lives and/or assets. 

• The significance of this work is that a knowledge of the plane where the largest concentration of radiation is located can help minimize loss of lives to fire hazards and improve the effectiveness of risk assessment and management.

• Developed and validated pool fire models.

• Used PHAST 7.2 to simulate data for view factor calculations from different fire scenarios. 

• Performed safety and hazard analysis of large pool fires for near and far afield observers.

• Performed hyperparameter tuning of pool fire models in comparison with PHAST 7.2. 

E. Other Master's Level Projects at The University of Bath, UK, Sept. 2014-15

• I modeled the performance a of two-channeled hemodialyzer with a single rectangular membrane using asymptotic analysis and numerical approximations. The resulting model has the potential to predict the efficiency of a Hemodialyzer to filter substances such as wastes and water from the blood when the kidney becomes dysfunctional. Hemodialysis helps to control blood pressure and balance important minerals, such as potassium, sodium, and calcium, in the blood.

• I investigated the design of a static mixer and modeled its mixing efficiency by using Computational Fluid Dynamics ANSYS FLUENT software packages. I carried out selective parametric studies of the static mixer's inlet diameter; mass flow rate; and inlet angle to determine their effects on mixing optimization. I found that regions with high turbulence kinetic energy yield greater mixing of fluids. This work is useful in the transportation of fluids such as chocolates, lotions, and sauces which are very common in food process and personal care industries.

• I modeled the carbonation process in concrete structures to determine the interface of carbonation depth. This is very useful in the design and reinforcement of concrete structures because carbonation causes cracks which can lead to the collapse of concrete structures. The modeling process takes into account the physico-chemical process of carbonation, i.e., the diffusion of atmospheric CO2 into a concrete slap and the consequent depletion of Ca(OH)2 seated within.

F. Bachelor's Degree Project at The University of Nigeria, July 2012

• I developed a mathematical model on the outflow of urine through the urethra, to study the transient behaviors of, and the relationship between the pressure from the bladder and velocity along the urethra during voiding. This project has a direct implication in our understanding of voiding dysfunctions, a condition that affects the ability of the body to urinate normally.