Inspiration
The average attention span for a Gen Z/ Millennial has been reduced to the length of a Tiktok, we wanted to stand out by mirroring the entertainment culture of today and integrating it with what makes baseball so magical, Homeruns.
What it does
Base fundamentally shows the user homeruns of their fave players and shows the analytics built on gemini.
How we built it
The database that serves the Homeruns is Chromadb base on The home runs datasets we were served, the were transformed using gemini's embedding models, enabling semantic search. Analytics are served by gemini, that analyzes each frame to return data like bat speed, the ball type and detailed explanations and descriptions. Gemini in this case uses the videos to identify the teams that were playing and players present in the video and shows the user. The play by play nuances are explained by gemini and are effectively given for noobs and nerds to understand an enrich their knowledge All this is powered by an underpowered Ubuntu virtual Machine on GCloud and we love it.
Challenges we ran into
The technology was fairly new and was very overwhelming at first but we then narrowed down to building with gemini. GCloud was also pretty new to us so it too some time to understand what we were doing but eventually we did
Accomplishments that we're proud of
How well home runs and its analytics run and are served.
What we learned
- Vector Databases.
- Cloud Console
- AI Studio
- Gemini Emeddings
- The difference between flash and pro
What's next for Base
Integrating Articles, Games and Standings Finally an AI Agent powered by vertex AI to talk to about Homeruns, Games And Standings
Log in or sign up for Devpost to join the conversation.