This is a website made for a conference workshop that took place on 4/25/24 to give the ~50 attendeeds hands-on experience and a basic understanding of prompt engineering. This web application was developed utilizing Typescript, React, Node.js, and OpenAI's GPT API.
As the conference has passed, the API is disabled.
How a struggle with training and contextualizing LLMs led to the development of a chrome extension to scrape and clean text for training data.
For training and contextualizing LLMs gathering, cleaning, and chunking text to reach specific token limits is a time consuming and error-prone process. A better solution is needed to reduce the time spent.
Defined a product goal and user stories for a chrome extension to solve the identified problem.
Designed a wireframe and mock-up of the chrome extension to visualize the user experience and features.
Translated and prioritized user stories into a backlog on Github projects to track progress, issues, and features in development. Also to support onboarding of collaborators.
Developed front end of the chrome extension using HTML, CSS, and JavaScript with all features. Will comment out features not yet developed, and slowly add as functionality is developed.
Prior to developing my own APIs utilized existing API/API keys to test functionality of the chrome extension. Verified json data was being returned.
Currently in development of GoLang API to handle data processing and cleaning of scraped data from Public APIs. Developed & tested a REST API to move jsons in Postman. Once that is complete, will be publishing limited functionality MVPs to gather user feedback for continued development.
Non-functional features commented out, leaving bare bone extension that converts website url to .txt output to be delivered to users for initial testing & feedback.
Garden Grove
California, US
© 2024 Joshua Gagnier. All rights reserved.