- Develop and implement audio-driven facial expression animator used for 3D avatars

- Develop and benchmark computer vision models to segment and classify chromosomes from metaphase images- Develop computer vision algorithms to automatically generate labelled training data using metaphase images and karyogram images- Develop and benchmark Pap Smear detection and classification models- Teach weekly, 2-hour seminar series on deep reinforcement learning

- Develop automated data collection and analysis system for strawberry and cannabis smart farming- Develop and benchmark computer vision models to detect and classify target crop and its plant state (growth stage, nutrient deficiency state, insect damage, etc.)- Develop and implement strawberry disease state estimation models for powdery mildew and gray mold- Develop predictive models for soft-sensors and a GUI application using Python and Kivy

- Design course materials and conduct lectures for Probability and Statistics for Civil Engineers (CE 311S)- Teach basics of probability and statistics: Probability Theory, Random Variables, Normal / Lognormal Distribution, Bernoulli / Poisson Processes, Monte Carlo Simulation, Confidence Intervals, Hypothesis Testing, Regressional Analysis, and Reliability-Based Design

- Design Bayesian network of reservoir properties variables to predict ultimate recovery factor- Statistical data analysis on comprehensive reservoir properties database- Correlate production performance with geologic model- 3-D reservoir simulation to history match and forecast on representative play areas- Assist with all reservoir and well performance inputs into the production model- Model heavy oil electromagnetic heating with ferromagnetic nanoparticles- Develop simulation models to determine stimulation efficiency and sensitivitySupervised by Dr. Larry W. Lake

- Performed phase behavior experiments to find optimum salinity and surfactant formulation for specific crude oils- Conducted alkaline-surfactant-polymer (ASP) core-flooding to evaluate oil recovery- Developed software to calculate polymer viscosity at different shear rates, temperatures and monovalent/divalent ion concentrationsSupervised by Dr. Gary A. Pope

- Develop workflow to automate adaptive coarsening of complex geological reservoir models to minimize simulation runtime while retaining key flow characteristics- Conduct adaptive upscaling in both vertical and areal directions to optimally find a reduced representation of geological features in the reservoir

- Develop probabilistic models for production rate forecasts and reserve estimates for every lease in Texas- Develop a model to determine and quality control the optimal subset of data that accurately represents the production behavior by lease- Help develop lease valuation model and PUD location model

- Studied gas hydrate dissociation behavior- Conducted literature review on production methods for hydrates- Helped researchers edit, proofread and publish their theses- Attended seminars and academies on alternative energy sources