~/talk

Minutes Matter: Fielding Edge AI for Personnel Recovery in 6 Weeks

links:none on file
links:none on file
slides:none on file

Built a TAK plugin that detects and labels buildings on overhead imagery, ships in six weeks with six people, and runs fully on aging S20/S23 Android phones and legacy Windows (.NET 4.x WPF) laptops. Team evaluates YOLO, Mask R-CNN and UNet variants; uses PyTorch -> ONNX edge runtimes and a custom .NET–Java JNI bridge. Result: GRGs auto-detected and auto-labeled in about two minutes (versus eight+ hours manually), editable by operators, meeting gov cyber requirements and offline constraints. runs fully offline

AI-written summary · about

  1. 0:00Introduction by Gene Kim
  2. 2:58Personnel recovery scenario
  3. 5:33Building the GRG
  4. 7:27Offline, aging devices constraint
  5. 10:56Building the edge AI stack
  6. 12:25Defining what a building is
  7. 17:44Engineering problems worth owning
  8. 21:24Asks and closing lessons
~/search
type to search 6,518 conferences, 143,918 speakers and 192,133 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
← →
on a talk: 5 seconds back / forward
0–9
on a talk: jump to 0%–90% of the recording
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.