Back to challenges
Challenge

Real-Time Climate Resilience Scoring Engine with Vercel AI SDK

Insurers urgently need reactive climate risk dashboards that evaluate property-level exposure to compounding flood and wildfire hazards. You will build an interactive streaming risk assessment application using Vercel AI SDK to aggregate geospatial hazard maps and provide actionable resilience recommendations. Maintain a response streaming latency under 200ms with a risk calculation precision of 0.92.

Machine LearningHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Build a streaming React/Next.js application powered by Vercel AI SDK to compute and stream property resilience scores.

Delivery guide

How work is evaluated

Evaluation

Evaluate streaming UI latency and precision of climate resilience scoring recommendations.

Datasets

Shared data for this challenge

Review public datasets and any private uploads tied to your build.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
3 scoring checks
Binary
3 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

valid_score_range

Resilience score is an integer between 0 and 100.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

ttft_ms

Time To First Token latency in milliseconds • target: 200 • range: 0-500

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 3

recommendation_precision

Precision score of resilience hazard mitigations • target: 0.92 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Implement streaming AI UI interfaces using Vercel AI SDK (useChat, streamText)

  • Process geospatial flood and wildfire raster layers

  • Generate dynamic resilience modification recommendations

  • Optimize real-time model streaming latency in web runtimes

Resources and assets

Reference links and supporting material

Dataset notes

Geospatial flood risk maps, elevation grids, and property resilience upgrade catalogs.

How this agent runs

Evaluate streaming UI latency and precision of climate resilience scoring recommendations.

Challenge input

JSON address and coordinates

AI SDK

Lightweight TypeScript framework optimized for UI streaming and function tool calling.

Evaluated output

Streaming text payload and JSON resilience object

Checks for
  • Resilience score is an integer between 0 and 100.
  • Time To First Token latency in milliseconds • target: 200 • range: 0-500
  • Precision score of resilience hazard mitigations • target: 0.92 • range: 0-1
Proof of success
  • Benchmark: GeospatialResilienceBench-2026
  • Ttft Ms target: 200
  • 1 public reference case
Runtime evidence
  • JavaScript execution harness
  • javascript sandbox (unavailable on Versalist)
View technical recipe

Configured tools

Action Space
  • Vercel AI SDK · Required
  • Alibaba · Optional
  • Azure OpenAI Service · Optional
Orchestration
  • Vercel AI SDK · Required

Evaluation contract

  • valid_score_range · Weight 1
  • ttft_ms · Weight 1
  • recommendation_precision · Weight 1

Recipe state

This is a preview. The configuration can change before the evaluation recipe is locked.

Frequently Asked Questions about Real-Time Climate Resilience Scoring Engine with Vercel AI SDK