Skip to main content
Loading navigation
Back to challenges

ALS Biomarker Progression Predictor with Vercel AI SDK

NIH discoveries show that novel blood-based biomarkers (such as neurofilament light chain, NfL) can predict when Amyotrophic Lateral Sclerosis (ALS) symptoms will emerge prior to clinical manifestation. In this challenge, developers will construct a web-based clinical dashboard backend using the Vercel AI SDK (TypeScript). The system will process patient serum biomarker concentration trajectories, call forecasting functions, and stream real-time onset predictions to clinical trial coordinators.

Data ScienceHosted by Vera
Sign in free to start

What you are building

Build a real-time streaming ALS biomarker progression predictor using the Vercel AI SDK and TypeScript.

How work is evaluated

Evaluation

Evaluates accuracy of estimated time-to-symptom-onset based on longitudinal biomarker inputs.

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

positive_onset_estimate

Predicted months must be a valid positive numeric value

Binary check

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

Dimension 2

onset_mae_months

Mean Absolute Error in predicted symptom onset months • target: 1.5 • range: 0-12

Binary check

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

Learning goals

  • Implement streaming AI endpoints using Vercel AI SDK (`ai` library) in TypeScript

  • Structure tool execution for calculating longitudinal biomarker trajectory slopes

  • Integrate clinical predictive schemas with Zod validation in TypeScript

  • Stream real-time disease onset probability estimates to web clients

Resources

Dataset notes

Synthetic plasma biomarker longitudinal readings (NfL, pNfH) paired with clinically verified symptom onset timestamps.

Challenge setup

Evaluates accuracy of estimated time-to-symptom-onset based on longitudinal biomarker inputs.

Challenge input

JSON containing patient_id, nfl_readings_pg_ml (list of objects with month and value)

AI SDK

Mandatory Vercel AI SDK

Evaluated output

JSON containing predicted_months_to_onset (number) and trajectory_slope (number)

Evaluation checks
  • Predicted months must be a valid positive numeric value
  • Mean Absolute Error in predicted symptom onset months • target: 1.5 • range: 0-12
Configured evaluation evidence
  • Benchmark: ALS-Progression-Bench
  • Onset Mae Months target: 1.5
  • 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
  • Merge · Optional
Policy Serving
  • Gemini 2.5 Flash · Optional
Orchestration
  • Vercel AI SDK · Required

Evaluation contract

  • positive_onset_estimate · Weight 1
  • onset_mae_months · Weight 1

Metrics

  • onset_mae_months · Target 1.5

Benchmarks

  • ALS-Progression-Bench
  • biotech

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing patient_id, nfl_readings_pg_ml (list of objects with month and value)

    Output: JSON containing predicted_months_to_onset (number) and trajectory_slope (number)

    Execution harness: JavaScript

Evaluation tools

  • AI SDK · Agent Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about ALS Biomarker Progression Predictor with Vercel AI SDK