Transforming Wildfire Management: Google Cloud Highlights WindTL’s Innovation
Last week, Improving Aviation had the incredible opportunity to host the Google production team at our office for a video shoot about WindTL, our wildfire prediction platform. The two-day event brought together our team, Google’s experts, and a great production crew to highlight the advancements we've achieved through our collaboration with Google.
Showcasing WindTL's Impact
WindTL tackles the long-standing challenge of delivering accurate fire spread and ember ignition predictions to those on the field. Embers are tiny, wind-carried firebrands responsible for 90% of structural losses during wildfires by igniting new fires miles ahead of the main blaze. WindTL predicts the spread and ignition of these embers by analyzing multiple conditions in real-time, including vegetation, topography, weather, and fire fronts, and delivers real-time fire spread and ember spotting using a highly intuitive, easy to use visualization platform. WindTL’s ability to analyze data in real-time and combine physics with machine learning makes this platform a leading solution to predict fire spread and ember ignited spot fires—an innovation that improves both preparedness and response efforts.
Google’s video highlights how WindTL works behind the scenes and in the field. The Google production team captured our technology in action, from data analysis on our backend systems to drones flying over controlled fires, collecting and processing real-time data. These scenes showcase the real-world impact of WindTL, demonstrating how it empowers firefighters to act decisively and insurance companies to better assess and mitigate wildfire risks.
A Partnership That Transformed WindTL
Improving Aviation was selected for Google for Startups' first-ever AI cohort earlier this year. Over the past eight months, the collaboration with Google’s coaches, mentors, and technology specialists has reshaped WindTL’s trajectory.
This partnership allowed us to evolve WindTL to tackle one of the most critical challenges in wildfire management: the delivery of accurate fire predictions to those on the field. Outdated prediction techniques have long failed to provide the actionable data in real-time that is needed to make timely decisions in the field.
With the computing power of Google Cloud and its capabilities to fuse and analyze data in real-time, we can now deliver actionable data in real-time. By leveraging advanced tools like Google Kubernetes Engine (GKE), we reduced run times from 30 minutes to just a few minutes, enabling multiple users to run WindTL in the field . Additionally, integrating Vertex AI and BigQuery has enhanced our model’s artificial intelligence capabilities, enabling continuous training and delivering improved accuracy over time.
What Our Team is Saying
The visit wasn’t just about showcasing our technology; it was also a moment of reflection and celebration for our team. Here’s what some of our employees had to say about the experience:
"One of the highlights of the video was the field work, where we demonstrated how drones collect data over a controlled fire. Watching the data seamlessly process through our backend systems in real-time was such a powerful moment—it really showed how WindTL bridges advanced technology with practical application."
"As a software engineer, this project pushed me beyond my usual technical role. It was an incredible opportunity to share the behind-the-scenes development of our model and reflect on the collective effort that makes WindTL such an invaluable tool. Being part of this process really reinforced the real-world impact of what we do"
Looking Ahead
This video marks a great moment for Improving Aviation, celebrating not just what we’ve built, but the people and purpose behind it. As we prepare for the release of the final video in the coming months, we are reminded of why we do what we do: intelligence for the most complex problems. Thank you for supporting our mission to break barriers and build resilience in wildfire management.