Skip to main content
Back to challenges
Challenge

Cross-Border PayNow-UPI Settlement Agent with Ax Optimization

Singapore and India have linked their fast payment systems (PayNow and UPI) to facilitate real-time, low-cost cross-border remittances. However, currency volatility, liquidity imbalances between settlement banks, and transaction fee thresholds require adaptive parameter tuning to maintain high success rates while minimizing operational liquidity costs. In this challenge, you will design an autonomous financial routing agent built on the OpenAI Agents SDK. The agent optimizes liquidity allocation and routing thresholds across multiple participating banks using Ax (Adaptive Experimentation). The agent must inspect incoming PayNow-UPI payment streams, dynamically adjust FX margin thresholds, and trigger automated balance rebalancing actions when liquidity limits are breached.

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 an autonomous cross-border payment settlement agent using OpenAI Agents SDK and Ax adaptive experimentation to optimize liquidity and routing.

Delivery guide

How work is evaluated

Evaluation

Evaluates agent routing efficiency, liquidity optimization accuracy, and compliance log generation.

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

settlement_success_threshold

Ensures settlement success rate exceeds 98%

Binary check

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

Dimension 2

liquidity_cost_savings

Percentage reduction in idle capital liquidity reserve requirements • target: 25 • range: 0-100

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 OpenAI Agents SDK for function routing and tool execution in financial workflows

  • Integrate Ax Bayesian optimization to dynamically tune payment routing parameters

  • Model real-time PayNow-UPI settlement queues and liquidity thresholds

  • Construct auditable risk logs for regulatory compliance under MAS standards

Resources and assets

Reference links and supporting material

Dataset notes

Simulated PayNow-UPI real-time payment log dataset with currency fluctuations and multi-bank liquidity pool states.

Challenge setup

Evaluates agent routing efficiency, liquidity optimization accuracy, and compliance log generation.

Challenge input

JSON containing payment transaction queue and bank liquidity state

OpenAI Agents SDK

Core requirement for autonomous agent logic and tool orchestration.

Ax

Required for Bayesian parameter optimization of payment routing.

Evaluated output

JSON containing agent decision, optimized routing parameters, and rebalance execution log

Evaluation checks
  • Ensures settlement success rate exceeds 98%
  • Percentage reduction in idle capital liquidity reserve requirements • target: 25 • range: 0-100
Configured evaluation evidence
  • Benchmark: FinTech-Bench-2025
  • Liquidity Cost Savings target: 25
  • 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

Action Space
  • Ax (Adaptive Experimentation) · Required
  • OpenAI · Optional
  • Vellum · Optional

Evaluation contract

  • settlement_success_threshold · Weight 1
  • liquidity_cost_savings · Weight 1

Metrics

  • liquidity_cost_savings · Target 25

Benchmarks

  • FinTech-Bench-2025
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing payment transaction queue and bank liquidity state

    Output: JSON containing agent decision, optimized routing parameters, and rebalance execution log

    Execution harness: Python

Evaluation tools

  • OpenAI Agents SDK · Agent Framework
  • Ax · Adaptive Experimentation

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Cross-Border PayNow-UPI Settlement Agent with Ax Optimization