I'm a physicist by training, but now I'm a data scientist. I’m a maestro of finding the golden needle of insight buried within the enormous haystacks of data. I love solving technical problems of any sort, but I’m especially interested in causal analysis, differential privacy, and machine learning. I’m an extrovert who doesn’t mind working directly with clients, and I have a knack for explaining complex topics in simple terms.
I have experience with a variety of data science tools, ranging from Spark to Pandas. I’m fluent in both Python and Matlab, and am comfortable with tools like Git or Linux. I’ve worked with natural language processing using SpaCy, with generative adversarial networks using TensorFlow, and of course basic regression and classification using tools like XGBoost.
Using publicly-available data about names from a Latin American country, I'm building a graph of family relationships. This will be useful for research, as well as identifying inappropriate business relationships between family members of government officials.
This repo is set to private, to protect PII, but I can add you if you're interested in seeing it.
Phonon measurements are a powerful check on the accuracy of theoretical calculations such as DFT (Density Functional Theory). The theory should calculate the positions between the electrons, and thus the forces between the atoms, which shows up in the phonon spectrum. This repo can be used to calculate the energies and intensities of the phonons according to the DFT predictions.
With large area detectors it's possible to collect scans from hundreds of Brillouin Zones in a single experiment. Except along high-symmetry directions, phonon measurements can be plagued by overlapping signals. But by refining multiple zones simultaneously, we can measure those enegries quite precisely, as I described here