Learning language
15.6 trillion tokens
Text at foundation-model scale. Meta’s Llama 3.1 405B was trained on a vast record of written language.
Llama 3 paper, 2024Visual AI for the built environment
Intelligence for
the built world.
We build AI to understand buildings: what they’re made of, how they change, and where they may fail. Starting with roofs and exteriors.
Roof materials
Identify surfaces and exterior features across a market.
Insurance · Lending & bonds · Manufacturing · Property
Building intelligence,
from site to portfolio.
Materials, wear, and exposure reveal the physical asset behind an address. We turn those details into intelligence for a property, a portfolio, or an entire market.
Explore building intelligence
01
Find the right buildings.
Identify properties by roof material, exterior features, and location. Build a focused view of a market or service territory.
02Understand their condition.
Recognize materials, locate visible defects, and track exterior change. Focus inspections and maintenance where they matter.
03Connect condition to risk.
Bring building characteristics and environmental exposure into underwriting, collateral review, and portfolio analysis.
The next learning problem
A world of images.
A missing dimension: time.
Language models learn from text. Robotics learns from action. We’re building the data to learn how the physical world changes: how a material weathers, a defect develops, and a building fails.
Read our research thesisLearning action
1 million+ trajectories
Open X-Embodiment combines experience across 22 robot types to study manipulation.
Open X-EmbodimentInside the technology
A world model for
how buildings change.
Failures unfold over years. We create virtual environments to study materials, weathering, and damage together—and generate the training data that’s hardest to collect.
Inside our approach Condition over time
Surface change adds condition information to the same building.
01 Simulate conditions
02 Train the model
03 Evaluate on real buildings
Industry applications
Physical insight.
Practical advantage.
The same building tells a different story to an insurer, a lender, or a manufacturer. Start with the decision your team needs to make.
Insurance & reinsurance
Bring the building into the underwriting decision.
Connect exterior condition and hazard exposure for property underwriting, renewals, and loss control.
- Underwriting & renewals
- Loss control
- Portfolio accumulation
Lending & real estate credit
Understand the physical assets behind the loan.
Add building condition and exposure to collateral review, capital planning, and portfolio monitoring.
- Collateral review
- Capital expenditure review
- Ongoing monitoring
Bonds, structured credit & surety
Connect financial exposure to physical assets.
Study the buildings behind asset-backed credit, bond portfolios, and surety exposure.
- Asset-backed credit
- Bond portfolio research
- Surety & project review
Manufacturers & roofing teams
Find the buildings your products are built for.
Identify roofs and exterior materials across a territory, then focus market development and outreach.
- Building discovery
- Territory planning
- Product & field research
Owners & facility teams
See a portfolio at the level of each building.
Organize exterior condition, maintenance priorities, and climate exposure across a property portfolio.
- Condition screening
- Maintenance planning
- Weather preparedness
Data, imagery & AI teams
Build the dataset your next model needs.
Collaborate on synthetic datasets, specialized vision models, and building-focused evaluation.
- Synthetic datasets
- Model development
- Data partnerships
Commercial underwriting In development
Your next deal.
A clearer picture.
Property details, deal assumptions, and physical risk—connected in one review.
Explore the productBuild with us
Tell us about
your project.
From a market or portfolio to a new model, we’ll help define the data and analysis your team needs.
Or reach us at rortega@rakeml.com
Start a conversation
Share a few details about your team and what you’re working on.