learning aws, mongodb atlas, and more in 24 hours—my sunhacks 2024 experience

never thought i’d learn so much in 24 hours—and i’m talking big stuff, like the tools used in modern software development. from aws to mongodb atlas, it was a wild ride, full of firsts.

at sunhacks 2024, i was responsible for the backend of memorylane—an ai-powered tool we built to help dementia patients relive their memories. and boy, was i thrown into the deep end.

using aws for the first time

before sunhacks, i’d heard people trolling the aws interface, but never really knew if it was that bad. turns out, it is that bad. setting up my first s3 bucket was a real challenge, and after spinning around in circles for a while, i decided to skip the docs and use llms as my guide. after about an hour, i managed to get the s3 bucket up and running (and felt like i deserved a medal). then came elastic beanstalk—i figured, since we were gunning for an aws prize, why not deploy the backend there? turns out, heroku was like training wheels compared to eb. who knew heroku was basically an aws wrapper, right? but eventually, i got the node.js backend running on eb too.

enter mongodb atlas

the next challenge was working with mongodb atlas, my first encounter with a nosql database. we used it to store the tags generated from google cloud vision api and the s3 file urls. configuring the cluster and integrating it with the backend wasn’t as straightforward as i expected, but once we got it going, it was pretty smooth.

my fourth hackathon—and what i learned

this was my fourth hackathon (but first at asu), and while we didn’t win, i learned more in those 24 hours than i have in some weeks of lectures. the feedback from the judges was incredibly motivating, especially from ojima abraham, who gave us some great advice on how to take this idea forward.

hot take: actually building something practical

what i loved most about this hackathon was that we built something that solves an actual problem. there’s nothing wrong with chatbots, chrome extensions, or the usual api plug-and-play, but i wanted something more practical. we may not have won, but i’m more proud of building memorylane than i would’ve been with a simple chatbot project.

the next step? who knows, maybe a hardware device…

at the end of the day, i walked away from sunhacks 2024 with more than just a project—i got hands-on experience with modern backend tools that i’m definitely using in the future.

check out our project on devpost: memorylane