Apr 2024 – Mar 2025 · 11 mos · Asheville, North Carolina, United States · Remote

• Reporting to an SVP, I led data science in the streaming Recommendations and Personalization space, building
• 1. Graph neural net solutions to solve the title cold start problem.
• 2. Time-series based solutions for Trending, Popularity and Top 10 algorithms.
• 3. Content embedding based solutions for Watch Next, Because You Watched, and You Might Also Like experiences.
• 4. Generative AI solutions to expand upon existing machine learning capabilities.
• 5. Reinforcement learning solutions to solve for page staleness.
• 6. Web application for model explainability so editors and product teams can better understand how our models work and what they specialize in serving.

Apr 2022 – Apr 2024 · 2 yrs · Asheville, North Carolina, United States · Remote

• I led data science in the streaming Recommendations and Personalization space, building
• 1. Graph neural net solutions to solve the title cold start problem.
• 2. Time-series based solutions for Trending, Popularity and Top 10 algorithms.
• 3. Content embedding based solutions for Binge, Because You Watched, and You Might Also Like experiences.
• 4. Web application for model explainability so editors and product teams can better understand how our models work and what they specialize in serving.

Aug 2020 – Apr 2022 · 1 yr 8 mos

• Focused heavily in Merchandising excellence, I worked to create machine learning solutions for freight optimization, space optimization, and assortment optimization, developing a world-class, user friendly localization tool that gave merchants a way to negotiate with vendors and make the best decisions on what to stock in every store. In addition, I created automated database connection utilities and an automated outlier detection mechanism for data scientists in the organization.

Jul 2019 – Aug 2020 · 1 yr 1 mo

• Having applied machine learning concepts to create powerful algorithms in the .com arena, I moved on to tackle brick and mortar modeling, starting up a small team and working with merchandising to provide assortment-based clustering and optimization, creating a world-class tool to provide localized recommendations based on customer interest and demand, explicit and implicit. We also worked with marketing to provide stronger insights to increase traffic and conversion for unique shoppers.

Feb 2017 – Jul 2019 · 2 yrs 5 mos

• Specializing in product assortment and personalization with a background in .com, I developed collection-selling models and dynamic models involving customer behavior and geographical trends to provide product and search recommendations.

Aug 2015 – Feb 2017 · 1 yr 6 mos

• Working with lowes.com Analytics, I helped with projects regarding personalization, experience testing, product variance selling, collection selling, and search engine optimization. I developed web crawlers in Python for QA purposes and designed many algorithms live on lowes.com, most involving collection selling logic.
• I've developed dynamic algorithms following customer behavior and aggregating for local and national trends, created algorithms utilizing search and product view baskets to provide common search results, and created a reporting suite using Hadoop and Alteryx with Location, Revenue, Orders, Clicks, Visits, and Impressions for all of the above items and all other Personalization algorithms.

Jan 2014 – Aug 2015 · 1 yr 7 mos · University of North Carolina at Chapel Hill

• The RESOLVE Galaxy Survey is focused on gathering kinematic data for a volume-limited sample of galaxies to more deeply understand the role of the environment on galaxies and what this environment is comprised of. My specific project this year involves comparing inclinations of galaxies as measured from kinematics and those as measured from photometry, and then exploring parameters such as size, mass, asymmetry, axis ratio, and more to see what photometric properties may correlate with differences in photometric and kinematic inclinations. After this work, I will be testing the robustness of the classical disk inclination formula, to see if any changes are required for more consistent matching.

Sep 2013 – Sep 2014 · 1 yr

• Data acquisition and simple conversions to help create the most extensive catalog of Gamma Ray Bursts in existence, to be made public when completed. I have completed the organization of approximately two years of data, or about 200 GRBs, each with 5-100 short papers. This work has contributed to the completion of all years up to about 2006. This work takes time and can be monotonous at times, but it felt great to be a part of such a massive undertaking.

Jun 2012 – Sep 2013 · 1 yr 3 mos

• I collaborated with Dr. Patrick Treuthardt in mapping out dust lane structure in relation to co-rotation radii and pattern speeds of galaxies using IRAF. I assisted in data collection in an open format, where museum patrons could watch it happening live on large monitors. I also helped in creating informative and exciting visual learning demos for the general public. Finally, we ran interactive H-Alpha "sun-gazing" talks, where the public could use a telescope to safely see the sun and learn more about it.