Trish O'Meara- Gillespie

Resume for Trish O’Meara- Gillespie

LinkedIn * **Technical Skills * JavaScript, TypeScript, SQL | React, Tailwind, Vite, Zustand, Vercel | Node, Express, Mongoose | MongoDB, PostgreSQL, Pinecone | Docker | OpenAI API, Claude, Gemini, Co-Pilot, ChatGPT | Git, GitHub, Linear, Postman, Jira, Figma *** **Experience *** Software Engineer | OSLabs | PatternBook - front end documentation generator

July 12, 2025 - April 19, 2026

• Designed with the promise of zero-config, PatternBook generates visual documentation in under 30 seconds. PatternBook requires no setup from the developer.
• Built with React and TypeScript to integrate seamlessly into existing developer workflows. TypeScript specifically was chosen to enforce type safety during documentation generation. This reduces errors across component trees.
• Patterbook’s UI is fully resizable and adapts to each developer’s workflow without causing layout constraints.
• Intentionally scoped to a stack that is popular with student and independent developers: TypeScript, React, Vite, and Tailwind, PatternBook makes codebases immediately approachable.
• Optimized for a minimal footprint, installing only necessary dependencies to reduce bundle size and accelerate developer workflow.
• Improves the speed of developer onboarding by generating visual documentation that maps component relationships, reducing the time needed to understand and onboard to a new codebase.
• PatternBook’s visual documentation is created by running a scanner on the components. This generates a JSON manifest that PatternBook consults and uses for documentation. This JSON file can be updated by running one or more of PatternBook’s commands. For example, “scan” creates or updates a manifest. The command “analyze” takes the manifest and shows it in the browser.

Production Manager

BCCO NYC

August 20, 2018 - January 21, 2025

• Coordinated and directed the daily operations schedule implementation, inventory and task delegation.
• Performed regular maintenance of brother GTX printers. Designed samples using Adobe Photoshop and corresponded with clients to ensure satisfaction with their order using MSOffice365.

• Researched and introduced the method known as “direct to film,” in-house saving 70% on price of materials and time. * **Projects *** Fake News | Full Stack Engineer

• Fake News is an app using AI and LLM for determining if a user provided statement is correct. The LLM is fed news articles as a JSON. Each entry on the JSON was modified to contain the necessary meta-data for processing with OpenAI’s “text-embedding-3-small,” this model takes text input and builds vectors.
• Fake News used Pinecone’s vector database to store the vectors generated. Utilizing Pinecone’s cosign similarity search to find and supply data that is potentially similar to the original user provided statement.
• Used a hashmap for memoization of repeated user inputs thus optimized time and space complexity. For example, popular inputs would trigger default messages repeating cached results.
• Re-refridge
• Re-refridge is an open source product developed to address the problem of how to reduce food waste. The front end UI allows a user to register an account or sign in and takes user input and stores it on a database. This database table is the user’s ‘fridge.’
• Users can create multiple fridges, add/ delete items to the fridge. The items all contain metadata with expiration or use by dates. It is also possible for two users to share a fridge.
• The app notifies users when their fridge has items that are going to expire.
  * **Education *** George Washington University Bachelor of Fine Arts May 2016

This site is open source. Improve this page.