Provided RLHF data to various large language models from major leading AI companies such as Google, OpenAI, and Anthropic, in accordance with many different guidelines that changed between projects. This data was created with the use of A-B comparisons between responses from large language models (LLMs) that gave preference to responses to user queries that were more factual, well-formatted, and more helpful overall based on a multi-dimensional axis system. Additionally, I assessed different capabilities of multimodal models that operate in software suites that use integrated LLM services, as well as services that assessed external software suites such as Google Sheets and Microsoft Office Excel, assessing their ability to pull and utilize data from their data sources. Finally, I gave technical feedback about the functionality and stability of different live services that also integrate these models.

Performed comparative ratings for search engine results for Google's web search service, optimizing results based on lengthy guidelines by the major tech giant. Additionally, comparative ratings were given for results for the large language model Gemini, providing RLHF data to help better train the model to provide well-structured, factual responses that are the most optimal for helping users in any scenario. Much of the work focused on web search results initially, but transitioned to RLHF data later on in my tenure with RaterLabs, up until the company’s last days

Trained over 40 team members in the kitchen as well as drive-thru and frontline positions, showing new members how to effectively serve customers with friendly communications, speed, and safe food handling in line with national food safety standards. Assisted with managerial tasks and worked closely with the owner and director of operations to ensure that new teams including new shift managers were onboarded effectively during periods of major team changes as turnover churned through employees both new and old.