Founding team

The team behind AlwaysHome

Built in one day for the UCLA Anderson MSBA Vibathon 2026, theme 04 “Uninsurable”: helping California homeowners on the FAIR Plan find their way back to affordable private insurance.

Portrait of David Ewenczyk

David Ewenczyk

Co-founder & Chief Data Officer

Architect of the data engine and the coverage model. Turned scattered state, federal and insurance data into one validated pipeline for Los Angeles County, and built the total-loss model that shows homeowners what their policy would pay, and what it wouldn't.

david.ewenczyk.2027@anderson.ucla.edu

MSBA Class of 2027, UCLA Anderson

Portrait of Joaquin Ramirez

Joaquin Ramirez

Co-founder & Chief Product Officer

Leads product and AI. Built the homeowner experience end to end: the AI policy reader that quotes the exact line behind every number, the house-damage model trained on 22,000 homes from the 2025 fires, and the platform it all runs on.

joaquin.elias.ramirez.gutierrez.2027@anderson.ucla.edu

MSBA Class of 2027, UCLA Anderson

Portrait of Atharva Uplap

Atharva Uplap

Co-founder & Chief Risk Officer

Leads wildfire risk science. Built the neighborhood wildfire exposure index from state hazard maps, 75 years of fire history and FEMA loss data, and the insurance gap and recommendations engine powering our next release.

atharva.uplap.2027@anderson.ucla.edu

MSBA Class of 2027, UCLA Anderson

Data

CAL FIRE / Office of the State Fire Marshal
Fire Hazard Severity Zones (SRA 2024, LRA 2025)
CAL FIRE
Historical fire perimeters since 1950; damage inspections (DINS) of the 2025 Eaton and Palisades fires
FEMA
National Risk Index: wildfire frequency and expected annual loss by census tract
California Department of Insurance
New, renewed and non-renewed policies by ZIP (2015 to 2023); FAIR Plan share by ZIP (2022); Safer from Wildfires measures
USDA Forest Service
Wildfire Risk to Communities
US Census Bureau
Geocoder, American Community Survey, ZIP and tract crosswalks (with HUD)
Zillow Research
Home values by ZIP

Built with

  • Claude API (Anthropic)
  • Next.js and React on Vercel
  • FastAPI and Python on Render
  • GeoPandas, pandas, scikit-learn
  • Streamlit (internal test bench)

Prototype for education. Not insurance advice. Fix costs and discounts are estimates; scores describe areas and homes like yours, not an inspection of your property.