The team
building spatial
intelligence.

Founded 2023 · San Francisco
37.1831°N   37.0422°E
ORIGIN

We started in the field — real-time rescue coordination during the 2023 Turkey earthquake. That experience showed us what spatial AI actually requires. We've been building it ever since.

Read our manifesto →

FIELD DISPATCH  /  TURKEY EARTHQUAKE  /  FEB 06 2023

“Real-time coordination
while rubble was
still settling.”

That deployment exposed the gap: no system existed to turn raw sensor data into queryable spatial understanding fast enough to matter. So we built one.

47
structures assessed

Turkey earthquake response, 2023

3h
assessment window

From raw sensor data to structured output

2023
founded

Berkeley, CA → San Francisco

Who we are.

01
Alper Gel, CEO

Alper Gel

CEO

Berkeley grad. Background in computer vision and disaster response. Led real-time coordination during the 2023 Turkey earthquake, which became the founding moment for Percept.

02
Victor Wu, CTO

Victor Wu

CTO

Dartmouth grad. Built end-to-end computer vision, geospatial analytics, and data processing systems. Northeastern MS in ML/SWE. Shipping the reconstruction and understanding pipeline.

03
Sam Cornelison, CGO

Sam Cornelison

CGO

Berkeley grad. 5+ years enterprise sales into public safety. Founding AE at Prepared and Hyper, got first annual contracts for both. Built our $1.2M pipeline in 3 states with $90K spent.

Guided by experts.

Decades of experience in emergency management, spatial computing, and public safety.

Dr. Thomas Azwell, Director, UC Berkeley Disaster Lab

Dr. Thomas Azwell

Director, UC Berkeley Disaster Lab
Dr. Joshua Dimon, Environmental Researcher, UC Berkeley Disaster Lab

Dr. Joshua Dimon

Environmental Researcher, UC Berkeley Disaster Lab
Rachel Stern, Founder & CEO, GovTech Ventures

Rachel Stern

Founder & CEO, GovTech Ventures
Dr. Heni Ben Amor, Associate Professor, ASU

Dr. Heni Ben Amor

Associate Professor, ASU
Dr. Hannah Kerner, Assistant Professor, ASU

Dr. Hannah Kerner

Assistant Professor, ASU

Make the physical
world programmable.

We believe that understanding the physical world should be as accessible as querying a database. Every building, road, tree, and terrain feature should be indexed, queryable, and simulatable.

Our technology reconstructs environments from any sensor, understands them at a semantic level, and simulates real-world physics — all through a single platform and API.

Read Our Manifesto →