Working at Broadcom this summer was a great experience that shaped my career growth and introduced me to a space where collaboration was at the forefront of the software we built. The work culture and sheer professionalism displayed by everyone working here were something that I admire. Broadcom employees were extremely receptive to helping newer members understand the environment we are in and how to leverage the opportunity. My team was nothing short of spectacular in the way they handled multiple projects with grace and made significant improvements every day. It was inspirational. The project we did was a Chrome extension inside of Broadcom’s mainframe support internal software. What this extension does is it helps support engineers have access to an AI-Powered Chatbot inside of their support software that can help summarize cases, make draft articles, help them submit lead forms, show them chart analytics of cases and who its assigned to, give them a summarized history of any internal comments that are inside of cases (this can be lengthy). Our goal for this extension was to help them resolve cases more efficiently, increase productivity, and reduce the time it takes to answer cases. Some of the technical stacks we used for the extension were: JavaScript, Python for backend hosting of our Flask server, Docker, which we used to package all our dependencies, Jenkins for our CI/CD pipelines automation, we used Google Cloud to track our project's resources, and Artifact was used to tag our Docker instances. GitHub was used to track the version control. Github was also utilized to merge our code by creating multiple branches. We used a terminal for running our builds, and also Chrome for developers is where the extension stays. Nginx was used to host our LLM server, and Google Drive was used to automatically update the different versions we rolled out to engineers.
I utilized JavaScript in the extension to create the front end of the extension, injecting JavaScript inside the chrome api while having different DOM Manipulation techniques to monitor Activity inside the extension. Using Complex Regex patterns, we were able to create a script that accurately flags any PII or sensitive information on the extension. In the overarching system of a job, when a user sends their message to the LLM API, this is then routed to the proxy server that is hosted on NGINX, then back to the frontend of the extension. The codebase is broken down into different modular sections, which include the scripts that handle all the logic for the Chrome extension, a build folder that contains the extension code that is given to users, a backend files that help with initializing all the routes for the Flask API. During the deployment of our Docker instance, the typical development process should be in the following order: implementing changes, testing a basic run by starting the backend locally, and then testing a Docker build with the associated commands. Finally, adding those changes to a feature branch of the GitHub repository and then opening a pull request to merge into the dev branch. Jenkins was utilized to automate the CI/CD pipeline in the extension and track different changes using Jenkins scripts to package all the dependencies for automation, and then Artifactory was used to tag our Docker instances while being under the MSD-support folder to keep track of versions and build summaries. For distributing the extension, we opted to use Google Drive to store and allow users to download the Extension.
This Internship educated me a lot about how to collaborate in a professional setting with other engineers, especially in certain areas like how to merge code and the best practices for product development. It informed me that there are multiple ways to overcome roadblocks in software engineering, especially given some of the restrictions and challenges we faced during production. It saddens me to depart from a software I've grown to love. Coming in the doors every day, knowing this is what I've always dreamed of as a career, loading up my laptop, and signing into where I was going to dedicate my hours, didn't feel like a job. I would stay day and night if it meant working more on this software. Seeing others around me compelled by their love for the work motivates me to give my all. Broadcom is more than its name: it's the commute, the connections, the love for engineering; it's having a dream, an idea, and executing it without looking back. It's being young and having ideas that people are receptive to. It's getting numerous coffees each morning without worrying about an upset stomach because you're fueled by the inspiration of the things you're creating. Broadcom will ring in my head like church bells at midnight, and I'll always listen, knowing my career started at the citadel where talent is bred. I've developed a deep tolerance for any problem I face and a stronger ability to problem-solve. Although this might not be a goodbye or a short period of growth, I'll always know that Broadcom is a place where the young become professionals and where the experienced wield a sword of knowledge thick enough to cut through any problem. Although I'll miss the long days and even longer nights working, I've left better than I was before.