Research Experience
Undergraduate Researcher
CurrentResearch Focus: TBA
Key Contributions: TBA
Undergraduate Researcher
CurrentResearch Focus: TBA
Key Contributions: TBA
Data Science Intern
Summer 2025Research Focus: Benchmarked U-Net vs. Neural Cellular Automata (NCA) for amodal segmentation on synthetic multi-view lab data, demonstrating that NCA achieved comparable IoU (0.46 vs. 0.25) while using 18% fewer parameters. Implemented a full model-to-amodal completion pipeline, converting 3D U-Net architectures into 3D convolutional video models and improving encoder-decoder RGB reconstruction across diverse comparisons.
Key Contributions: Led experimental design and analysis for model comparison, including model-to-amodal inference, RGB binarization, and multi-modal training. Analyzed model generalization by comparing RLS-based estimators with data-driven predictors, informing ongoing work on hybrid ML + control pipelines for autonomous driving systems.
Undergraduate Researcher — Pandey Group
CurrentResearch Focus: Developed an online Recursive Least Squares (RLS) estimator for real-time parameter identification in Adaptive Cruise Control (ACC) systems, achieving < 0.05 steady-state error across highway, suburban, and stop-and-go simulations.
Key Contributions: Designed and analyzed a model-to-amodal completion pipeline to test control algorithms under varying driving conditions, enhancing system reliability and adaptability. Improved short-horizon velocity forecasting by 18%. Analyzed model generalization by comparing RLS-based estimators with data-driven predictors, informing ongoing work on hybrid ML + control pipelines for autonomous driving systems.
Undergraduate Researcher — Kim Group
Summer 2024Research Focus: Developed and analyzed computational simulations of the Susceptible-Infected-Recovered (SIR) model to study epidemiological trends for COVID-19.
Key Contributions: Implemented Kinetic Monte Carlo (KMC) to model stochastic disease transmission within an agent-based framework on lattice structures. Designed bash scripts for parameter sensitivity analysis. Processed and visualized data using NumPy and Matplotlib.
Personal Projects
Ecological Time Series Population Predictive Model
Developed a data-driven ecology model using the Lotka-Volterra equations to simulate rabbit-bobcat population dynamics. Fitted parameters to real sighting data via least-squares optimization and projected wildlife population trends. Preprocessed and interpolated sparse ecological time-series to a common grid, normalized data to avoid scale bias, and produced stable input for modeling despite noise and incomplete observations.
Impact: Generated population forecasts under varying predator-to-prey ratios to predict whether the bobcat population would exceed an ecological threshold (200 individuals), providing actionable insight for wildlife-management decisions.
Presentations & Posters
Benchmarking U-Net and Neural Cellular Automata (NCA) for Amodal Segmentation
Presenters: Axel Muñiz Tello, Melanie Gomez-Bustos, Wesley Hur, Eddie Lizarzaburu
Investigated the application of computer vision techniques to synthetic data for training models aimed at automating laboratory processes. Integrated simulation-based datasets with machine learning methods to enhance the reliability and scalability of automated experimental workflows, demonstrating the potential of synthetic training data to accelerate innovation in automated scientific research environments.
Modeling Disease Dynamics Using Kinetic Monte Carlo
Presenter: Axel Muñiz Tello
Explored disease transmission dynamics using Kinetic Monte Carlo (KMC) simulations within an agent-based SIR framework. Modeled populations on lattice structures with restricted and unrestricted movement, examining how spatial structure and stochastic effects shape epidemic behavior and offering insights for improving future disease control strategies.