Vibing in the Wild: Algo Astra and the Developer That Didn’t Do Hackathons

I want to start by saying this project was never meant to happen. In all honesty, I had developed an intense bias over my career in software against hackathons in general. In my eyes, they were not worth the extra effort on top of my daily developer responsibilities. Additionally, there was always an underlying feeling that no matter how useful my idea was, how hard I worked, or how well I presented my solution, I would never receive full credit for my efforts. 

Parts of this disillusionment began to dissipate on the way to Jackson, Mississippi for the MagnoliaJS conference. This was my first time traveling to Mississippi, but well beforehand, I’d heard of the wonderful organizers behind Magnolia, and how hard they worked to unify, welcome, and create opportunities for all levels of tech professionals. In addition to the ride with my tour buddies, Calypso, and Alex, Day 0 concluded with dinner hosted by the conference organizers. In between jokes about the latest AWS outage and new tech buzzwords, we bonded over southern favorites like fried green tomatoes and Mississippi distilled whisky flights.

Day 1 of the conference centered on a hackathon for attendees. I originally had no plans to participate, but after such a warm and friendly welcome, I wanted to be where the action and the fun was. The rules of the hackathon were as follows:

1. Create a project (AI and vibe coding were allowed) 

2. The project must include at least ten different components from a corporate sponsor’s library 

3. The project needed to be ready to present in 6 hours

The project could be anything we wanted, and that type of broadness terrified me at first. 

While I have professional React experience, due to the demands of my most recent roles, my skills had grown stale. Still, before thinking about routes and components, I had to reach a decision on what I could build with the given rules. Several concepts began floating around my mind, so I turned to chatGPT to help organize my development plan. I didn’t have time to write a full design/product requirements document, but on reflection, my first semester of graduate school as a software design and architecture student taught me better. ChatGPT took my ideas and suggested an educational app that combined my interests in data structures and algorithms and my passion for teaching people of all ages how to code. Furthermore, a list of features and the potential library components that could be used to build them was generated.

I was off to a magnificent start, but trouble was on the horizon. This was great and helpful information, but now I had to take it to an IDE and start writing code. After ensuring my initial setup with Vite and React were in place, and that necessary packages were installed, I tried copying a section of code directly from my chat. At first glance, the code syntax was correct and linted properly, and the React hook it wanted to implement also followed all the necessary rules. There were no compilation errors, either. Yet, my browser displayed a blank screen in place of my component, and a 403 forbidden network error that I later realized was due to trying to access API key shielded components in the library. While this was a rookie mistake, I was definitely glad that I had the previous experience and knowledge to understand what went wrong. I also realized that I would need a better pair programming partner. 

Enter Claude. A few weeks prior to Magnolia, I had the opportunity to attend a Claude coding workshop presented by the expert engineers at This Dot Labs and led by my friend and VP of Engineering, A.D. Slaton. Illuminating is an understatement. The presenters walked us through error debugging, getting Claude fully set up and integrated with our Github accounts, and how to better organize and clarify prompts to build functional components. By the conclusion of the workshop, we had a fully implemented codebase written with help from Claude. While reviewing my new repository afterwards, I felt how I imagined the first humans to use flint to start a fire might have felt.

Back at the hackathon, after updating my Claude credentials, I kicked into high gear, recalling how to get all of my API connectors configured. Next, I started prompting. This time, I was careful to ensure that I included the documentation for accessing the free component library. I also had a lot less time to work with at this point, as over half the competition was complete. Claude met me exactly where I needed, and I began generating the necessary parts of my repository. This time, the example React fragment worked. I spent the remaining time adding more features to cover more advanced data structure topics, and quizzes for users to test their knowledge.

The single page app that I presented was far from pretty, and the confidence I had in my skills was tested. I was so proud that I rose to the challenges I encountered, and that mattered more to me than prizes or recognition. I was officially a software developer again. I was reminded that my ideas mattered, and that my skills accumulated thus far across my journey in software mattered as well. 

Long before this hackathon, I had realized that the traditional technical individual contributor role was never created for someone like me in mind, at least not for long term. I was, always have been, and once again reminded myself that I am a builder and tinkerer. Though my experiences and rejections in corporate made me feel lesser about my abilities, they could never take away what I knew, the work I built, or what I was capable of.

I want to specifically express gratitude to the organizers, speakers, and community at MagnoliaJS for reminding and affirming this to me as well. 

When I returned home, I kept building. Eventually I gave my project an official name, a color scheme, and is now available for the public here, AlgoAstra. Future planned enhancements include the ability to have user profiles, saving and averaging previous quiz scores to give users a more accurate sense of where they may need improvement, and expanding topics to include my almost 200 page encyclopedia of software engineer interview preparation materials. Ultimately, my hope for Algo Astra is that learners enjoy interacting and practicing their skills even more than I loved building this project.