George Mitchell | Junior Analyst | Mathematics PhD

George Mitchell

Junior Analyst | Mathematics PhD

Based in New York, USA

About

Mathematics PhD graduate from CUNY with research experience at Microsoft. I enjoy solving all kinds of problems, from the theoretical to the practical, using tools from math, statistics, computer science and data.

Skills

Projects

BloomDash

An interactive dashboard for building a stock portfolio and analyzing the returns. Built using Python, Dash and Plotly, the dashboard allows for plotting of returns, extracting alpha and beta of the portfolio against the S&P500, getting buy and sell signals, and calculating the Sharpe ratio.

Predicting Loan Default

We predict loan default based on over 10 years of historical loan data. We use Polars for data handling, Optuna for parameter tuning and Seaborn for visualization.

Forecasting Stock Volatility

We forecast the volatility for a bundle of US equities using various time series models, including: Random Forest, GARCH and a LSTM neural network. This work was completed as a final group project for the Erdos Institute Data Science Bootcamp, May 2023.

Experience

2023-Present

Junior Analyst at Zink Inc

Creating dashboards and reports to summarize real estate data. Building linear models to describe and predict real estate metrics.

2022-2022

Research Intern at Microsoft

Improving the Square-Root Velu algorithm for evaluating isogenies between supersingular elliptic curves. Generalized the biquadratic relations to any number of points on an elliptic curve.

2023-2023

Coach at Modal Learning

Coaching courses in machine learning, data science and python programming.

Education

2012-2016

Mathematics Degree at Newcastle University

2016-2022

Mathematics PhD at City University of New York

2022

Data Science Bootcamp

Intensive bootcamp hosted by the Erdos Institute, aimed at graduate students in STEM fields. Bootcamp covered all aspects of supervised and unsupervised machine learning, including time series forecasting and deep learning.