Brandon O'Briant

DATA SCIENTIST


Inspiration:
Core Values:
Background
Vision:
Interests include: Systems engineering, data science, deep learning, machine learning, computational modeling/simulation, telecommunications/networking engineering, expert decision support systems, optimization/automation, sensing systems, cyber-physical security, agricultural automation, and high performance computing.

Ph.D. Systems Engineering


About Colorado State University's Ph.D. in Systems Engineering Program

M.S. Data Science


About Northwestern University's MSDS Program

Northwestern University's M.S. Data Science Program (MSDS) provided the opportunity to hone and apply critical skills to real world problems. Areas of study and application include relational database systems and analytics software, traditional statistics and machine learning methods.

multivariate analysis and generalized linear models, text knowledge discovery and analysis, interactive visualization with D3.js, mathematics of risk, probability/statistical theory, linear and non-linear integer programming, goal programming, metaheuristic algorithms, and Markov decision models.


Course Work


Courses completed, courses scheduled to be taken, and notable, real-world projects conducted in the MSDS program are listed below. The links take you to Brandon's GitHub repositories associated with the course. His repositories are continuously being updated with new projects and the refactoring of previous projects, so check back frequently for updates.

Highlights of Coursework


Personal Projects


Linear Regression with Gradient Descent (C++)

Currently, I am working on building a Machine Learning library in C++ from scratch. This linear regression is the beggining of my journey. Takes in x and y values, performs linear regression using gradient descent. Each iteration is printed to the console with the final result is displayed at the end. There are several features I am currently working to add (most likely after finals in a couple weeks): Function to take in data from file, train-test-split with KFold Crossvalidation, evaluation metrics, and CUDA optmization.

Markowitz Portfolio Optimization

With new tech companies emerging everyday, which four companies should I build my Tech Portfolio with? Given those four choosen, how can I optimize this portfolio to maximize returns, while minimizing total cost? Python was used to create an optimized portfolio of tech securities Apple Inc (AAPL), Microsoft (MSFT), Intel (INTL), and Tesla Inc. (TSLA) by running 1000 different weighted possibilites to produce the Markowitz Efficient Frontier.

Capital Asset Pricing Model (CAPM)

Sharpe then introduces the market portfolio concept. This market portfolio is constructed as a combination of all possible investments in the world (bonds and stocks). This allows for bundling of securities in such a way that maximizes risk-return, while optimizing (minimizing) risk via diversification. Thus, this portfolio will lie on some point on the efficient frontier and is the most efficient portfolio. The Sharpe Ratio is calculated and visualized using Python.


Skills & Tools


Brandon has developed proficiency and expertise in the following programming languages and comfort with the following tools, while continuously expanding his technical abilities.


Contact


Brandon is interested in expanding the network of professional acquaintances, hearing about potential career opportunities, and welcomes thoughts or questions about his work:

obrinabl@protonmail.com
CV