video1.0<iframe src="https://www.loom.com/embed/02d8cd6855844d119fcd969e2a7aa2ad" frameborder="0" width="1818" height="1363" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>13631818Loomhttps://www.loom.com13631818https://cdn.loom.com/sessions/thumbnails/02d8cd6855844d119fcd969e2a7aa2ad-d762b810fdc6fd83.gif192.341This Loom demo explains how Sightline helps triage critical resources during citywide storms in real time. It describes aggregating reports such as 911 and 311 calls, drone sightings, social media posts, and utility data to determine where support should be dispatched most urgently. The zero-shot model achieves a fit of 60.7, and within the initial rescue critical field it reaches about 70 out of 101, emphasizing the importance of accuracy in life-or-death situations. It then moves to Harness Evolution, which trains on historical storm sightings and iteratively improves and validates performance, with a live arena showing Jev predictions running in real time.