Engineering Decisions Without Field Visits

Diving deeper into

Orest Pilskalns, CEO of Skyfish, on building autonomous drone infrastructure

Interview
you can make the engineering decisions without having to go back into the field and verify
Analyzed 5 sources

This is the shift from drones as camera tools to drones as engineering systems. When Skyfish can turn flight data into a measured 3D model that identifies parts, aligns multiple sensors, and stays accurate enough for tower and utility workflows, the expensive step of sending a crew back out just to confirm dimensions starts to disappear. That moves the product closer to system of record software, not just airborne data collection.

  • In practice, the savings come from collapsing a two step workflow into one. Instead of flying a site, building a rough model, then dispatching another crew with tape measures or specialty tools, engineers can inspect the model in Skyportal, measure components, and make design or maintenance calls from the office.
  • The key technical unlock is not only rendering a mesh, it is knowing what each object is and where each frame was captured in space and time. Skyfish ties together custom hardware, sensor timing, geotagging, and 3D machine learning so images from different sensors line up cleanly enough for engineering use, not just visual review.
  • This puts Skyfish in a different lane from software first mapping vendors like DroneDeploy, which make generic drone data usable across many workflows, and closer to a precision inspection stack for telecom and power. Within domestic drones, Skyfish is grouped with Wingtra and Freefly in high end engineering and inspection rather than public safety autonomy led by Skydio.

The next step is that accurate models become the starting point for recurring software revenue and eventually autonomous inspection networks. As more tower, utility, and power workflows trust office based measurement, the winning drone companies will be the ones that own the data pipeline from sensor capture to model to decision, then plug that into drone nest operations at scale.