Projects
Sockeye Salmon Forecasting
Alaska is among the most food insecure states in the US. In response I developed a forecasting model to predict the 1-day ahead Sockeye Salmon inflow to the Kasilof River using a Kolmogorov Arnold Network (KAN). This involved procuring, cleaning, engineering features, and training over 300 KAN models using a stochastic model building process. This is one of the first projects to asses the viability of KAN networks in an ecological setting that compared it to alternatives like multi-layer-perceptrons (MLPs). The KAN based model demonstrated tangible predictive power that could potentially increase subsistence harvest via a mobile or web based app. To learn more click on the buttons below.​​
Pricing Iowa Crop Insurance
Soy and corn farmers face substantial financial risks due to fluctuations in future prices and adverse weather events like hail storms. This project aims to provide an actuarially fair insurance policy for farmers while ensuring an insurer remains solvent during periods of catastrophic loss. Historical data was leveraged using Stochastic Diffusion Processes, Iman-Conover like procedures, a Historical Copula binding approach and Gaussian Kernel Smoothing to obtain simulated scenarios across an insurer's (BOB). The techniques used in this project could be adapted to manage uncertainty in other domains such as portfolio construction or niche insurance products. To learn more check out the code base below.
Data Center & AI Microchip Analytics
During the summer of 2023 I built and presented a Power BI dashboard to executives at Intel detailing the performance of AI microchip products. Building the dashboard required cleaning, preparing, modifying and merging enormous excel files into a data structure compatible with Power BI. Transforming the data into visualizations appeared easy at first, but given the thousands of possible chart, graph, and style combinations I had to quickly develop a keen eye for user experience to narrow them down. The project culminated in a presentation detailing the interworkings of the dashboard to ~30 sales executives around the world via Microsoft Teams.
Remote Mineral Exploration
During the Summer of 2022 I worked with a geophysics team in a remote camp outside Nome, Alaska to assess the quality of a mineral deposit for a large mining conglomerate. My role required me to quickly pick up a variety of skills such as firearm handling, troubleshooting satellite internet uplinks, riding in a helicopter, plumbing, geophysics, and operating an all- terrain vehicle. Our team was also vulnerable to extreme weather events such as blizzards, and dangerous animals like bears. This experience enhanced my ability to work on a variety of tasks while staying composed under difficult working conditions.
Asset Manager Competitive Research
During my internship at Denali Advisors, the largest Native American owned quantitative based money manager in the US, I was tasked with developing a coherent picture of emerging trends within the tax-exempt institutional marketplace.This involved cleaning and transforming Pension and Investments survey data with Excel, Power Query, Power BI, and Python to create 30 years worth of panel data. Pertinent insights were presented to executives within Denali Advisors and the Alaska Permanent Fund Corporation via a Power BI dashboard. This dashboard will aid portfolio managers in maintaining a competitive roster of products amid their complex operating environment.
CatScript Programming Language
Building a programming language from scratch is a great way to understand programming at an intrinsic level. I collaborated with several team members and faculty at my college to develop a scripting language called CatScript. This involved developing a tokenizer, parser and evaluation engine using recursive descent. The language was capable of compilation in Bytecode and Java using the Java Virtual Machine (JVM). Software engineering principles like Test Driven Development (TDD) were used to manage complexity. Insights gained from this project could be applied towards optimizing the speed of time-sensitive work flows or developing niche programming languages for specific business needs. To learn more check out the code base below!
Financial Engineering Club
Given the mind boggling number of career paths available to financial engineering students at Montana State University my team and I developed an organization to connect talented financial engineering majors with educational and professional development opportunities to facilitate a coherent picture of life after graduation. Events included guest speakers from large corporations like Deloitte and Goldman Sachs as well as from local entities including Atrium Data Science and Next Frontier Capital. Our organization also offered social networking events where other financial engineering majors could meet each other.
E-Commerce Store Database
I Developed a fictional e-commerce store using SQLite, Redis, Velocity, and Java. This involved writing SQLite queries for fictional customers, invoices, employees, and various products that would pull data from the SQLite tutorial Chinook DB and display it in an interactive GUI using the Velocity web framework. After achieving basic functionality, queries with the potential for high web traffic were sped up using Redis caching.The skills gained through this project could be adapted to other business needs like integrating a database with Microsoft products like Excel or it could also be used to organize/restructure data.
