jst.kevv / student internship recommendation system
dtctn_01 // details conf: 0.97
student internship recommendation system preview

student internship recommendation system

github

honestly, this project was kinda ambitious in a sense that i’ve over estimated how my team and i can do in a span of 3 months.

this was a collective project from multiple course subjects, and one the subject is called decision support system.

if you were ever take this course, you would know that the main goal of this subject is to create a system that can help users make informed decisions based on data analysis—it’s basically a lot of math and statistics.

usually, in programming, we could do manual calculations based on what algorithm we want to implement. i originally thought we have to do like manual calculations as well, but then in the powerpoint presentation—our lecturer said that we can use machine learning or ai-based approaches to do the calculations for us.

it was going well at first—i built the machine learning algorithm using random forest algorithm (and a lot of synthetic data) and it was kinda giving a good result, wrapping it up in a fastapi and it all seems well and done.

but all good things must come to an end, and the end came in the most unexpected way.

you may be thinking that it maybe because of the machine learning part or the integration part. i wish that was the hard part but no, it was actually the website, the design, and how i manage the team. i miss calculated my team thinking that they were “good” enough that i can trust them with any of the work but sadly, no, i was wrong, and it was a hard lesson to learn.

don’t get me wrong, we did finish the project—kinda, and i don’t hate my team for that. looking back, maybe i should’ve gone with just a simple design and using reusable components instead of going with a custom design—but hey, you live and learn right?

[0.224, 0.628, 0.179, 0.730]