Samba Emergency Response Agents Crew

Context Diagram
Agent Flow Diagram
Detailed Agent Flow Diagram
Wildfire event, high risk and safe zones
Satellite Image
Satellite Image Analysis using vision model - llama90b
Human in the loop
Human in the loop (console)
Agents Interacting With Each Other 2
Agents Interacting With Each Other 1
Agents Calling Tools

Inspiration

Inspired by SambaNova's cutting-edge RDU technology, our platform ensures lightning-fast processing of real-time events, enabling immediate and effective disaster response.

What it does

The Emergency Response Agents is an agentic real-time emergency response platform, currently focused on wildfire management and evacuation coordination. The platform is highly adaptable and scalable, making it applicable to a wide range of natural disasters and emergencies, such as hurricanes, earthquakes, and floods. Its core objective is to deliver instant, accurate, and actionable guidance to rescue teams and civilians, enabling them to respond effectively during critical moments when lives and property are at stake.

The platform's ability to provide real-time situational awareness is pivotal in emergency scenarios. By leveraging SambaNova's advanced RDU-powered systems instead of traditional GPU-based architectures, it achieves unprecedented speed and efficiency in processing emergency data. Through the integration of the SambaNova API, the CrewAI framework, and NASA's Earth Observatory Natural Event Tracking (EONET) API, the system processes live data on wildfire locations, magnitudes, weather conditions, and spread patterns. This data is then synthesized to offer critical situational insights. The use of SambaNova's Llama 3.2 vision models further enhances its capabilities, enabling precise analysis of satellite imagery to track wildfire progression and assess impacts. With a structured data architecture that includes evacuation routes, safe zones, and high-risk areas, the platform ensures timely, data-driven decision-making. Its modular design and feedback loops enable continuous refinement of assessments, making it a robust and adaptive tool for managing emergencies.

GenAI Models Used

SambaNova's
- Llama 3.2 vision - 11b, 90b
- Llama 3.1 - 8b, 70b, 405b
VertexAI's (for comparing performance with SambaNova's models)
- Llama 3.2 vision - 11b, 90b
- Llama 3.1 - 8b, 70b, 405b

What's next for Samba Emergency Response Agents Crew

Samba Emergency Response Agents Crew is highly scalable. As part of the hackathon, it is currently performing well with wildfire events, and this use case can be expanded to address any disaster-related scenarios.