John Pangas Email and Phone Number

I embarked on an exciting research project in collaboration with Mozilla and the University of Calgary. This research focuses on developing a tool that leverages GPT-4, a Large Language Model, to automate the generation of test cases for the Firefox QA team. Designed to aid in the creation of test plans for new features in the Mozilla Firefox browser, we aim to publish a paper on this research soon.The tool operates by taking a few test cases, the feature, its description, and the test scope and feeding them into the Large Language Model. The model then generates test cases and ,using cosine similarity, the tool compares the test cases developed by the QA team with those generated by the LLM, identifying which cases are similar and which were missed. The QA team can then adopt, restructure, or include the generated test cases as needed.I also collaborate on the Mozilla Bugbug project, which automates various tasks on the Mozilla Bugzilla platform, where all bugs of Mozilla products are filed. Check it out https://github.com/mozilla/bugbugMy Mozilla Profile: https://people.mozilla.org/p/jpangas

I worked with Mozilla Firefox under the Outreachy program sponsored by the Software Freedom Conservancy. At Mozilla, I worked on the BugBug project, an open-source initiative aimed at enhancing the bug identification process using Machine Learning.Key Achievements:• Developed and deployed a machine learning model to automatically detect and tag accessibility-related bugs on Bugzilla (Mozilla’s Issue Tracker), streamlining issue identification and prioritization for improved developer efficiency.• Built a machine learning model to automatically detect and hide spam comments on Bugzilla, reducing noise and enhancing developer experience.• Integrated Bugbug's model training process with Weights & Biases for better model interpretability and training metric comparisons.• Facilitated model training on TaskCluster (Mozilla's CI platform) to provide greater resource flexibility and enhance CI-integrated model visualizations.

Medici con l'Africa Cuamm is an Italian non government organization with an aim of improving health care in various African countries. My role was to analyse and filter company data in search for important insights and different objects of interest.