Data Science Researcher
- Designed and implemented a cutting-edge autoencoder algorithm using PyTorch, enhancing protoplanetary disk image reproduction by 2-3 orders of magnitude, surpassing the accuracy of traditional MCMC algorithms.
- Engineered and automated a comprehensive data, machine learning, and scoring pipeline, decreasing processing time by 50% and boosting model optimization efficiency.
- Integrated the Weight and Biases (wandb) API to streamline performance and efficiency tracking, enabling real-time monitoring and rapid iteration.
- Applied MCMC algorithms to generate high-quality training and testing datasets, showcasing expertise in statistical modeling and synthetic data creation.
- Presented complex data analysis findings to academic audiences and co-authored research papers, enhancing scientific communication and collaborative research skills.