Skip to main content
Loading navigation
Back to challenges

Coral Bleaching & Marine Asset Risk API using Vercel AI SDK and Streaming SST Anomalies

Record-breaking ocean temperatures during El Niño trigger widespread coral reef bleaching, causing catastrophic losses for coastal tourism, marine aquaculture, and natural wave barrier insurance policies across the tropics. Marine asset insurers require real-time streaming APIs to track ocean heat content and alert policyholders before mass bleaching occurs. In this challenge, you will implement a web-native API using Vercel's AI SDK in TypeScript. The system will stream real-time analysis of NOAA Coral Reef Watch sea surface temperature (SST) anomaly feeds directly to client dashboards. Your solution will execute tool calling for spatial SST indexing, calculate Degree Heating Weeks (DHW), and stream structured risk assessments with low latency.

Machine LearningHosted by Vera
Sign in free to start

What you are building

Build a web API using Vercel AI SDK to stream real-time ocean temperature anomalies and coral bleaching marine asset risk.

How work is evaluated

Evaluation

Evaluates the accuracy of Degree Heating Weeks (DHW) calculations and TypeScript stream response structure.

Datasets

Public datasets, plus the private datasets you have access to.

How submissions are scored

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

zod_schema_validation

Streamed object complies strictly with Zod definition

Binary check

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

Dimension 2

dhw_accuracy

Accuracy of Degree Heating Weeks calculation • target: 0.99 • 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

  • Deploy TypeScript agent streams using Vercel AI SDK (`streamText`, `generateObject`).

  • Process spatial ocean heat content and Degree Heating Weeks (DHW) metrics.

  • Implement multi-tool JavaScript/TypeScript functions for environmental data retrieval.

  • Stream risk evaluations with real-time UI component rendering capabilities.

Resources

Dataset notes

Synthetic time-series array of sea surface temperature (SST) anomalies across tropical reef assets.

Challenge setup

Evaluates the accuracy of Degree Heating Weeks (DHW) calculations and TypeScript stream response structure.

Challenge input

JSON with reef_id, weekly_sst_anomalies_c

AI SDK

Mandatory TypeScript framework for streaming AI responses.

Evaluated output

JSON with dhw_index, alert_level, asset_risk_score

Evaluation checks
  • Streamed object complies strictly with Zod definition
  • Accuracy of Degree Heating Weeks calculation • target: 0.99 • range: 0-1
Configured evaluation evidence
  • Benchmark: MarineHeatBench
  • Dhw Accuracy target: 0.99
  • 1 public reference case
Configured runtime
  • JavaScript execution harness
  • javascript sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Action Space
  • Vercel AI SDK · Required
  • VESSL AI · Optional
  • Llama 3.2 · Optional
Orchestration
  • Vercel AI SDK · Required

Evaluation contract

  • zod_schema_validation · Weight 1
  • dhw_accuracy · Weight 1

Metrics

  • dhw_accuracy · Target 0.99

Benchmarks

  • MarineHeatBench
  • energy

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON with reef_id, weekly_sst_anomalies_c

    Output: JSON with dhw_index, alert_level, asset_risk_score

    Execution harness: JavaScript

Evaluation tools

  • AI SDK · Web Agent Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Coral Bleaching & Marine Asset Risk API using Vercel AI SDK and Streaming SST Anomalies