Level of Projects Being Developed

Gemini API Developer Competition
1

I’m a student and would like to understand the complexity of the projects being developed for this hackathon. Are they more basic, such as PDF chatbots or recommendation/filtering systems, or are they more advanced, like Devin AI or automated DevOps solutions, etc.?

2

Hey there and welcome to the community!

We have no idea who’s joined or what they’re developing. When it comes to competitions like this, it’s typically a range of varying skill levels.

It’s pretty much whatever you as an individual can get done in 3 months. That answer is going to be different for everyone.

3 Likes
3

Let me tell you based on the project I’m working on rn, it’s way more complex than just a basic pdf chatbox, and probably even more complex than Devin AI which was a scam in the end.

1 Like
4

Welcome to the community!

First and most importantly - the competition feels like it is meant to push all of us to try new things. So I would suggest that no matter what you think needs to be done - most of all you should be using this opportunity to try something new. This is a learning opportunity for all of us who are competing!

Keep in mind some of the criteria, including:

  • Is the submission surprising to those that are well-versed in Large Language Models (“LLM”)? (maximum 5 points)
  • Is the submission surprising to those that are not well-versed in LLM? (maximum 5 points)
  • Does the submission differ from existing, well known, applications in functionality? (maximum 5 points)
  • Does the submission differ from existing, well known, applications in user experience? (maximum 5 points)
  • Is the submission implemented through the use of creative problem-solving approaches? (maximum 5 points)

To me, this suggests that the following kinds of things would get very few points for all of those:

  • Simple classification
  • Trying to answer questions just from the LLM
  • Simple RAG (or even simple question answering with the large context)
  • Simple code / SQL generation
  • “Rewrite this as…”
  • “Generate ideas based on…”

I think just about anything that will get a higher score needs to be multi-step agentic, demonstrate something more advanced, or demonstrate something that has been seen before (PDF chatbots) that have a new or novel twist to them (how they present the results, or what is done with the results, for eample).

Again, however, most important is what you get out of it.
Being able to show that to the judges will go a long way.

Good luck!

2 Likes
5

I’m eager to know more about your project after the competition…