San Diego, California, United States
• Tackled various projects supporting business analytics team in producing insights on consumer satisfaction. • Demonstrated ability to acquire and apply new skills quickly by learning Rust from scratch and translating entire data processing codebase from C# to Rust. Implemented machine learning models into Rust application. • Improved quality of transformer text classification deep-learning models through Semi-Supervised Learning and Natural Language Processing (NLP). Performed research to create Unsupervised machine learning text classification models.
• Enhanced a convolutional neural network (CNN) to analyze telescope data and identify rare galaxies with dual active galactic nuclei (two centric black holes). Created synthetic and augmented data to expand dataset and enhance effectiveness of CNN. Fine-tuned hyperparameters through optimization techniques. • Reduced model by just under 90% its original size, resulting in faster prediction speeds and significantly improved performance.
• Tasked by Systems & Optics team to prototype and train an AI image regression model to autofocus image samples from microscopes, potentially reducing footprint/cost of company's genomic sequencing hardware. • Demonstrated problem-solving skills by creating a custom convolutional neural network (CNN) model architecture that outperforms MobileNetV3, ResNet 50 pre-trained models in accommodating larger 16-bit images. • Wrote functions to perform custom preprocessing/data augmentation, including random cropping and rescaling to maximize effectiveness of a given dataset for the CNN.
UC San Diego Naviaux Lab - Mitochondrial Research • Created 3D visualization model comparing the metabolic networks of related mitochondrial diseases, which helped scientists interpret and classify their evolutionary differences. • Assisted in metabolic taxonomy research by developing programs to better organize data and spreadsheets. • Advanced stalled projects by demonstrating analytical, troubleshooting, and debugging skills during yearlong internship. Demonstrated ability to respond positively and effectively to coding feedback.
• Designed and programmed 3D video game for the company’s suite of gaze-controlled video games played using the Tobii Eye Tracker. • Created “Ring Leader,” a flying-themed 3D game that tasks attention-challenged children (such as those with autism and ADHD) to fix their gaze on an airplane in order to move it through rings. “Ring Leader” has received positive player reviews. • Worked in a collaborative, start-up environment; contributed game concept ideas; wrote design and test documents; tested builds and fixed bugs; worked with UX/graphic designers on visual assets; took initiative to create original music when suitable music for the game was hard to find. • Demonstrated ability to make an impact by learning new programming languages and technologies quickly.