I develop deep generative models that can render photorealistic avatars with natural facial expressions and body gestures to power seamless interactions with Openstream.ai's Enterprise Virtual Assistants.

I lead a team that researches, develops, and deploys the GraphSLAM algorithm — an algorithm that uses pose graph optimization to generate a more accurate representation of the race car's environment and state estimate from noisy sensor data. As part of the Carnegie Autonomous Racing team, we showcased the first Formula Student race car to complete a fully autonomous lap in North American history at the May 2023 Formula Hybrid Competition at New Hampshire Motor Speedway.

I helped lead the Controls team in designing systems that determine the optimal trajectory and corresponding actuator commands for a Formula Student race car traveling at high velocities.

In the Homeland Sensors and Analytics Group, I experimented using cutting-edge Neural Radiance Field (NeRF) algorithms for rendering 3D photo-realistic scenes with limited image data from various data modalities. I learned more about the NeRF model architecture, how to fine-tune such machine learning models, and how to utilize the lab's superclusters to optimize their training process for improving computational time and performance.

I refined a Python web-crawling algorithm to collect more metadata attributes for millions of apps on the Google Play Store in an effort to collect and analyze data pertaining to app privacy label research at Carnegie Mellon Chimps Lab. I also learned how to parse extremely convoluted HTML algorithms and deploy and run experiments on Google Cloud.

I design and refactor Python algorithms that run computer simulations on cyber criminal behavior under the guidance of Carnegie Mellon Professor Larry Heimann. My work involves employing rigorous unit-testing protocols and the Continuous Integration (CI) software development process as well.

Loop is a risk technology insurance company that works to contextualize where it’s unsafe to drive. As an intern, I engineered a synthetic population model of Texas to locate Loop’s target customers for strategic marketing campaigns. I built the model using an Iterative Proportional Fitting algorithm that was based on various demographics and other factors pertinent to Loop’s rating engine. In addition, I also aggregated and performed data analysis of risk scores for roads in various census tracts in Texas using Pandas.