GRIT is a statistical, climate-oriented risk assessment tool for insurers and reinsurers, emphasising explainability and accuracy along with aggregation of big data for real-time statistical predictions.
-
Prototype was built in 24 hours using Lovable and Google Cloud for 3rd MIT x OpenAI Global Hack-Nation hackathon.
-
Placed 3rd (VC track) out of >2800 participants, with more details viewable on Hack-Nation page and LinkedIn.
-
Watch 60-sec demo video
NOTE: Demo is of prototype submitted to hackathon on 11/2025 - currently it is outdated and does not show the more-advanced functionality listed below.
(Some features may still be WIP)
GRIT is a real-time climate risk assessment tool for insurance, reinsurance, and catastrophe-modelling teams. It delivers sub-10m geospatial analysis with feeds from over 15 satellites, <5 second response time from all APIs, and over 100 parameters for payout scenarios, providing reliable, explainable, data-driven insurance estimates.
Designed to answer the core underwriting question: What are the climate-related financial risks to insure someone in this area, both historically and now?
- 37-point concentric grid per query (1 centre + 3 rings of 6, 12, and 18 nodes)
- 4 hazard composites (flood, wildfire, storm, drought) with overall risk score
- <5s second average API response time
- 10m resolution using Sentinel-1/2 data via Copernicus API
- Exponential severity scaling (
riskFactor^1.2) for loss estimation - Serverless auto-scaling via Supabase Edge Functions + Lovable Cloud