Skip to main content
Back to challenges
Challenge

Build an Automated Information Integrity Audit Pipeline with LangChain and Modal

Construct a scalable, multi-modal content verification pipeline for regional media monitoring. Using LangChain for stateful workflow execution and Modal for serverless GPU inference, your workflow will analyze digital media streams, verify factual claims, and flag deepfakes efficiently.

Workflow AutomationHosted 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 serverless LangChain workflow running on Modal GPU infrastructure to analyze public media integrity and detect misattribution.

Delivery guide

How work is evaluated

Evaluation

Evaluate accuracy of claim verification and detection speed on Modal infrastructure.

Datasets

Challenge datasets

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

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

trust_score_in_range

Trust score calculated between 0.0 and 1.0

Binary check

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

Dimension 2

verification_precision

Precision of classified claim statuses • target: 0.88 • 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

  • Build a LangChain multi-step chain for claims extraction and media auditing

  • Deploy custom audio/image deepfake models onto Modal serverless GPUs

  • Integrate vector search indices for rapid fact-checking against trusted sources

  • Output standardized trust scores for digital news publications

Resources and assets

Reference links and supporting material

Dataset notes

Curated collection of regional news articles with human-annotated factual claims and misattribution flags.

Challenge setup

Evaluate accuracy of claim verification and detection speed on Modal infrastructure.

Challenge input

JSON containing article_text and article_id

LangChain

Popular agent framework for stateful task chaining.

Modal

Provides serverless model execution infrastructure.

Evaluated output

JSON containing claims_list with fact_check_status and trust_score

Evaluation checks
  • Trust score calculated between 0.0 and 1.0
  • Precision of classified claim statuses • target: 0.88 • range: 0-1
Configured evaluation evidence
  • Benchmark: FactCheck_Bench_v2
  • Verification Precision target: 0.88
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

Environment
  • Modal · Optional
Action Space
  • LangChain · Required
  • Langchain · Optional
Policy Serving
  • Modal · Optional
Orchestration
  • LangChain · Required
  • Langchain · Optional

Evaluation contract

  • trust_score_in_range · Weight 1
  • verification_precision · Weight 1

Metrics

  • verification_precision · Target 0.88

Benchmarks

  • FactCheck_Bench_v2
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing article_text and article_id

    Output: JSON containing claims_list with fact_check_status and trust_score

    Execution harness: Python

Evaluation tools

  • LangChain · Agent Orchestrator
  • Modal · Serverless GPU Runtime

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Build an Automated Information Integrity Audit Pipeline with LangChain and Modal