Inspiration
After interviewing the police, emergency medical & fire-fight staff personnel at the BitCamp event. The main problem was data access and friction in decision-making. Such fieldOps teams have lot of data that is scattered in their repositories. If I could train a custom GPT with their data, it could help them with answers to queries vs links. Helping teams with efficiency (low search time) and improved decision-making. Thus helping them serve to save lives and property.
What it does
Fetches/summerizes mission-critical information within seconds. Supports multi-lingual queries. Provide real-time traffic maps all in one place!
How we built it
We have used: LangChain, OpenAI GPT APIs, Azure App service, Astro (UI), Streamlit
Challenges we ran into
Gathering data was a huge pain-point, as there was no dataset available for such a recent event.
Accomplishments that we're proud of
A RAG-based system to answer within context. Renders real-time traffic maps. Supports multilingual languages for natural language based queries.
What we learned
Azure Copilot, Azure AI search.
What's next for BuildKey AI
Handle Images and video to get insights.
Log in or sign up for Devpost to join the conversation.