Geospatial Physical AI · The intelligence layer for the physical world.
Rayford AI builds Ray, a geospatial physical AI system for auditable property and infrastructure intelligence across disaster, resilience, and real-world operations.
Ray turns remote sensing, street-level imagery, and geospatial context into auditable intelligence for disaster, infrastructure, and property decisions.
Start with disaster. Build toward physical-world intelligence.
Ray is an intelligence system for real-world perception, evidence, and action — beginning with the most urgent workflow in property decisions: what changed after a disaster, and what it means.
| Module | Focus |
|---|---|
| Ray Assess | Property-level damage evidence — compare pre/post imagery, score visible damage, attach evidence, and surface confidence at the parcel level for auditable assessment workflows. |
| Ray Claims | Claims and inspection triage — rank properties for adjuster review and package imagery, metadata, and explanations for faster, more defensible insurance claim workflows. |
| Ray Risk | Hazard context and mitigation intelligence — connect damage evidence with hazard context, mitigation priorities, and resilience planning for properties and critical infrastructure. |
- Link world context — Parcel records, hazard context, built-environment data, and pre-event imagery assembled for the target area.
- Compare multimodal evidence — Street-view, satellite, drone, and field imagery across time, fused and aligned at the property level.
- Arbitrate model signals — Damage scoring, multimodal reasoning, and confidence estimates via Ray's CLIP-enhanced arbitration layer.
- Export audit trail — Action-ready evidence packages with scores, confidence, and imagery for downstream human review.
Ray's engine translates peer-reviewed GeoAI research into a practical intelligence layer.
- IGARSS 2026 — Satellite-to-Street: Synthesizing Post-Disaster Views from Satellite Imagery
- 2026 Preprint — DamageArbiter: CLIP-Enhanced Multimodal Arbitration for Hurricane Damage Assessment
- Computers, Environment and Urban Systems 2025 — Hyperlocal Disaster Damage Assessment Using Bi-temporal Street-view Imagery
- ICC 2025 · Best Student Paper — DisasterVLP: Perceiving Multidimensional Disaster Damages via Visual-Language Models
- Applied Sciences 2024 — GeoLocator: A Location-Integrated Large Multimodal Model for Geo-Privacy Inference
- Esri Press — Object Detection and Segmentation Using Text SAM in ArcGIS Online
Public research context is available at AutoGeoAI4Sci.
| Yifan Yang — Founder & Technical Lead | Multimodal spatial intelligence, street-view analysis, model arbitration, and autonomous GeoAI systems. |
| Dr. Lei Zou — Scientific & Technical Advisor | GeoAI and spatial-intelligence foundation; disaster resilience direction. Texas A&M. |
| Dr. Zhengzhong Tu — Technical Advisor | Computer vision, multimodal model design, and validation strategy. |
| Dr. Heng Cai — Technical Advisor | Built environment context, infrastructure intelligence, and product-risk review. |
This organization separates public company material from private venture work.
- Public repositories may include company profiles, website-facing assets, public documentation, and selected research links.
- Private repositories hold product code, internal strategy, customer discovery, model workflows, data pipelines, and other proprietary work.
Rayford AI is not an open-source project by default. Public materials are shared for visibility, evaluation, and collaboration conversations; proprietary code, data, designs, and venture materials remain protected unless explicitly released.
- Website: rayford-ai.com
- GitHub: Rayford-AI
- LinkedIn: Rayford AI
- Contact: contact@rayford-ai.com
Every property, ready and recoverable.
