Datawave (germanSolo)

Inspiration

Help SMBs in southamerica to be more datadriven.

What it does

Datawave aims to streamline this process, providing a centralized platform where teams can effortlessly synchronize, organize, and gain insights from various data artifacts without any coding knowledge.

How we built it

We used a streamlit as UI, Google cloud data tools like cloud SQL (relational database dummy) Cloud run (serverless Procesing power for contenerized apps) vertex AI (AI models) and a code in python using LLM libraries like pandasAI and langchain for rag and agents building.

Challenges we ran into

Video management was painfull because its hard to implement FFMPEG in a serverless enviroment we did a workaround using a AWS lambda function.

PandasAI is not friendly with gemini, we change to openai model for this tool of the app.

Accomplishments that we're proud of

The app works awesome! and it has ci/cd deployment with serverless google cloud tools.

What we learned

Multimedia management for llm based apps its hard because it has its own heavy dependencies

What's next for Datawave

Keep adding more conectors and go to Sales

Built With

• cloudrun
• cloudsql
• ffmpeg
• gemini
• google-cloud
• langchain
• openai