← Financial Cloud Cloud Cloud Club · AWS Re:cap

Vertex Macro | Financial Cloud Cloud · AWS Re:cap

AWS Re:cap 07: FSI Meetup 2025 Q4 - Financial Transaction Data Reconciler PayPal

Speaker: Jayaseelan Shanmugam

Session: 07

Session
Summit Dev Lounge2026 Re:cap
01 Encode Architecture as Steering for AI Agents
Summit Dev Lounge2026 Re:cap
02 Agent Harness Is the Real Engineering Moat
Summit Dev Lounge2026 Re:cap
03 Ask Observability Data in Plain Language
Summit Dev Lounge2026 Re:cap
04 Serverless AR Game with Bedrock AgentCore
Summit Dev Lounge2026 Re:cap
05 Multi-Agent Quant Backtesting on AgentCore
Summit Dev Lounge2026 Re:cap
06 Blog to Slides in Three Minutes with Kiro
Summit Dev Lounge2026 Re:cap
AWS Community Day Hong Kong 2025 Re:cap
02 AWS Compliance with Terraform
AWS Community Day Hong Kong 2025 Re:cap
03 Beginner to Builder An AWeSome Cloud Journey
AWS Community Day Hong Kong 2025 Re:cap
04 Team-First Serverless Engineering with Laravel & Bref
AWS Community Day Hong Kong 2025 Re:cap
05 Event Opening - AWS Community Day Hong Kong 2025
AWS Community Day Hong Kong 2025 Re:cap
06 Agent-to-Agent: Building Interoperable AI on AWS
AWS Community Day Hong Kong 2025 Re:cap
07 Utilize another telemetry data for faster improvement with AI agent
AWS Community Day Hong Kong 2025 Re:cap
08 Graduating from Vibe Coding: Spec-Driven Development with Kiro
AWS Community Day Hong Kong 2025 Re:cap
09 Automated Testing using MCP & AI Agents
AWS Community Day Hong Kong 2025 Re:cap
10 Modernizing Telecom Security ML Powered Approach
AWS Community Day Hong Kong 2025 Re:cap
11 Rethinking GenAI Agent: RAG & MCP
AWS Community Day Hong Kong 2025 Re:cap
12 Disaster and Emergency Response with TAK and AWS
AWS Community Day Hong Kong 2025 Re:cap
13 Rethinking Serverless Application Workflows from a Testing Perspective
AWS Community Day Hong Kong 2025 Re:cap
14 Practical AWS FinOps for Cloud Success
AWS Community Day Hong Kong 2025 Re:cap
15 AI-Powered Global Pure-Alpha Macro Trades on AWS: Revolutionizing Risk-Adjusted Asset Returns
AWS Community Day Hong Kong 2025 Re:cap
FSI Recap
01 Modern Trade Lifecycle: Trading to Settlement
FSI Recap
02 Goldman Sachs: Fast Track your applications onto Cloud - AWS Re:cap Q1/2023
FSI Recap
03 Zurich Insurance Group: Building an Effective Log Management Solution on AWS
FSI Recap
04 FSI Meetup 2025 Q4 - Brex Database Disaster Recovery
FSI Recap
05 FSI Meetup 2025 Q4 - A Graviton Migration Success Story
FSI Recap
06 FSI Meetup 2025 Q4 - Stifel Modern Data Platform
FSI Recap
07 FSI Meetup 2025 Q4 - Financial Transaction Data Reconciler PayPal
FSI Recap
08 FSI Meetup 2025 Q4 - Scaling Resilience
FSI Recap
09 Maximizing AI Inference Cost Efficiency: Strategic Adoption of AWS GPU Instances
FSI Recap
10 Advanced Agentic AI Design Patterns
FSI Recap
11 Build New Modern Apps on AWS
FSI Recap
AWS re:Invent 2025
01 Coinbase re:Invent Recap (IND3312)
AWS re:Invent 2025
02 Building the Future Trading Platform Leveraging AI and AWS
AWS re:Invent 2025
03 Trading Innovation: Jefferies' AI Assistant on Amazon Bedrock (IND3315)
AWS re:Invent 2025
04 How FSI Revolutionized HFT Analytics with Agentic AI (GBL302)
AWS re:Invent 2025
05 Improving Distributed Systems with Amazon Time Sync Featuring Nasdaq
AWS re:Invent 2025
06 Amazon Aurora HA and DR Design Patterns for Global Resilience (DAT442)
AWS re:Invent 2025
07 Building Agentic AI: Amazon Nova Act and Strands Agents in Practice (DEV327)
AWS re:Invent 2025
08 Deep Dive into Amazon Aurora and Its Innovations (DAT441)
AWS re:Invent 2025
09 Deep dive on Amazon S3 (STG407)
AWS re:Invent 2025
10 Nasdaq: Build Resilient Infrastructure for Global Financial Services (HMC327)
AWS re:Invent 2025
11 What's New with AWS Lambda (CNS376)
AWS re:Invent 2025
12 Spec-Driven Development with Kiro (DEV314)
AWS re:Invent 2025
13 Amazon's finops: Cloud cost lessons from a global e-commerce giant (AMZ308)
AWS re:Invent 2025
14 Tick to trade latency trading platforms on aws
AWS re:Invent 2025
Government data
01 The AI Era: The Boundary Between Development and Design Is Disappearing
Government data
02 On-Device Multimodal AI and Smart-City Practice
Government data
03 Large-Model Capability Evaluation and a Method for Landing AI Projects
Government data
04 Controlled End-to-End Automation of Government Development with Cloud Agents
Government data
05 A New Software Ecosystem for the Agent Era, Seen Through Multi-Agent Systems
Government data
06 AI-Driven Macro Quantitative Research and Smart Governance
Government data
07 Authorized Operation of Public Data and Smart-Government Practice
Government data
08 Putting Data Assetization into Practice: Rights, Compliance, Engineering Governance, and Digital-Government Cases
Government data
09 AI for Mental-Health Public Welfare: Governance, Architecture, and Practice of a Trustworthy Platform
Government data
Amarathon 2025 Recap
01 A Developer’s Roadmap to Architecting for Agents
Donnie Prakoso
02 Amazon Bedrock Data Automation
Hafiz Syed Ashir Hassan
03 Multi-Agent on AgentCore
Tan Xin
04 Building Agentic AI Nova Act and Strands Agents in Practice
Haowen Huang
04 Accelerating Migration Projects with Kiro using Spec-Driven Development
Sanchit Dilip Jain
06 From Matching to Understanding: Personalized AI Search Practice Driven by AgentCore Memory
Liu Cao
07 Observe to Optimize – LLM Observability to AIOps Turning real-time insights into intelligent automation
Jimmy Soh
08 Deploying TEAM and Building the Best Engineering Team
Yuji Oshima
09 Five Hard Lessons from Five Years of So-Called Serverless Databases
Renato Losio
14 What if AI does my job How Q Developer CLI and Kiro have changed my daily routine
Miguel Angel Muñoz
16 Velocity with Vigilance: Security Essentials for Amazon Bedrock Agent Development
Brian Tarbox
26 Run OSS LLMs on a Single H100 Smarter, Cheaper, Faster
Adit Modi
28 A Modern Unified Metadata Architecture: New Approaches to Breaking Down Data Silos
Shaofeng Shi
29 Serverless MediaOps: Automating Video Workflows with AI on Amazon Web Services
Luis Valdivia
30 Architecting for Efficiency and Reliability with Performance Testing at Scale
Luis Guirigay
31 Connecting the World Through Open Source: Practical Journey of Technology, Community and Global Developer Relations
Richard Lin
33 Building Streaming Iceberg Tables for Real-Time Logistics Analytics
Fahad Shah
34 Accelerating Large-Scale Robot Strategy Training: An Automated Closed-Loop Architecture Based on Kiro, Trainium, and EKS
Junjie Tang
35 From Vibe to Viable with spec driven development
Ricardo Sueiras
36 Making Cloud Cost Analysis Smarter: Building FinOps Intelligent Agents with Strands and AgentCore
Xiaofei Li
37 Transform Conversational Agentic AIOps for K8s Using CNCF Kagent, K8sGPT, and Nova Sonic
Shaoyi Li

Introduction to PayPal:

● PayPal is a global payment service provider processing 1.7 trillion in annual payment volume.

● Operates in 200 global markets with 430 million active accounts.

● Processes approximately 900 transactions per second, with peaks during holiday seasons like Black Friday and Cyber Monday.

Critical Problem:

● Ensuring the accuracy and reconciliation of the massive transaction volume.

● Reconciling transactions across multiple systems within PayPal, with external processors, and networks.

● Matching transactions to ensure no data or financial discrepancies.

● Validating that financial records accurately reflect actual customer payments.

Reconciliation Process:

● Transactions flow through PayPal’s system and are recorded in multiple internal ledgers.

● Transactions are sent to external processors for clearing and confirmation by the network.

● PayPal settles the transaction money to the merchant.

● Reconciliation involves matching transactions across PayPal’s internal systems, processor acknowledgments, and funding settlement summaries (end of day, T+1, T+2).

● Primary goal: Ensure transactions are not lost and there are no discrepancies.

Why Reconciliation Matters:

● Three-way matching problem: PayPal internal ledger, external processor records, and network confirmations.

● Critical for financial accuracy and customer trust.

● Ensures that financial records reflect actual customer payments.

High-Level Architecture:

● Focusing on how PayPal achieved near real-time reconciliation.

● Technologies and strategies used to handle the scale and complexity of the problem.

Business Impact:

● Reduction in reconciliation time from 24 hours to 15 minutes.

● Improved accuracy with minimal discrepancies.

● Enhanced customer trust and operational efficiency.


Continued Discussion on PayPal’s Reconciliation System

Three-Way Matching Problem:

PayPal Internal Ledger:

● Records transactions within PayPal’s systems.

External Processor:

● Registers transactions in their local system.

Network Confirmation:

● Confirms whether the transaction has been successfully made.

Responsibilities:

● PayPal is responsible for matching transactions at every stage from entry into the system to settlement with the merchant.

● Manual reconciliation is impractical due to the high volume of transactions.

Automated State Machine with Rule Engine:

● Utilizes an automated state machine and high-level rule engine.

● Configured to handle transaction processing with external vendors and timelines for acknowledgments and funding summaries.

● Ensures transactions are reconciled efficiently.

Importance of Reconciliation:

● Critical for understanding "what happened versus what actually happened."

● Ensures transactions are auditable and compliant with regulatory standards (PCI DSS).

● Provides a clear record of when transactions were recorded and settled.

Current Gaps in Legacy System:

● The legacy system relies on end-of-day batch processing.

● Uses a store and process mechanism where transactions accumulate throughout the day and are reconciled at the end of the day.

● Source of truth is not the direct operational data store to avoid performance and latency impacts.

● Utilizes an ETL system sourcing data from Oracle GoldenGate.

● ETL pipeline involves transformation and formatting, leading to potential data mismatches or inconsistencies.

Need for Improvement:

● Move away from batch processing to near real-time processing.

● Reduce reliance on ETL systems to minimize data transformation issues.

● Enhance automation to ensure accurate and timely reconciliation.


Problems with Legacy System:

● Experiences delays due to accommodating all transactions until the end of the day.

● Transactions are matched and account books are closed only at the end of the day.

● This delay is a significant problem.

Objective:

● Transition to a new age platform in the cloud.

● Leverage AWS infrastructure to solve the aforementioned problems.

Key Objectives of the New Solution:

End-to-End Data Integrity Across Payment Lifecycle:

● Ensure data integrity from the moment a record enters the real-time payment processing system.

● Track transactions across multiple systems within PayPal.

● Link all transactions with the correct identifier and timestamp.

● Match outbound files sent to processors with inbound records received from networks or vendors.

Automated State-Driven Match Logic:

● Move from a store-and-process mechanism to a stream-and-process mechanism.

● Reduce the entire reconciliation cycle.

Real-Time Monitoring:

● Identify exceptions while matching transactions within the internal system or records from external vendors.

● Record exceptions where matches fail between received records and local ledger transactions.

● Operational team to act on these recorded exceptions.

Technical Architecture:

Data Injection:

● Sources from which data is injected into the reconciler.

Reconciliation Process:

● Methods and processes involved in performing reconciliation.

Storage of Reconciliation Outcomes:

● How the results of the reconciliation are stored.

Operational Team Leverage:

● How the operational team uses reconciliation exceptions and acts on them.


High-Level Technical Reconciliation Overview

Scope Confinement:

● Upstream payment processing systems are abstracted out.

● Focus starts with the real-time payment card processor.

Real-Time Payment Card Processor:

● Utilizes EKS service to receive millions of transactions per day (expected ~300 million transactions daily).

● Each transaction is recorded in AWS DynamoDB, which serves as the source of truth and operational data store.

Data Flow:

● [ 1 ] DynamoDB to Kinesis Data Stream:

● Transactions recorded in DynamoDB are streamed via Kinesis Data Stream.

● Kinesis manages ordering of transactions.

● [ 2 ] Amazon Data Firehose:

● Transactions are bucketed and chunked based on different business parameters.

● [ 3 ] AWS S3:

● Transactions are recorded in AWS S3.

● S3 acts as a secondary data store for transactions but primary for file processing.

Reconciliation Process:

● [ 1 ] Inbound Transactions in PayPal:

● Processed transactions in PayPal are translated into file format in AWS S3.

● [ 2 ] External Partner Processing:

● Chunk files are translated into files and processed with external partners.

● Inbound records from external partners are returned to S3.

AWS S3 as Central Source:

● S3 holds both internal transaction footprints and external processed transactions received as inbound files.

Data Processing:

● EventBridge Scheduling:

● Triggers Apache Spark on AWS EMR cluster every 15 minutes.

● Distributed processing of transactions in S3 using a preconfigured rule engine.

Rule Engine:

● Comprises multiple state graphs.

● Categorizes transactions and determines terminal states.

● Includes complex rules based on market operations, external partners, and cutoff times for data export/import.


Technical Reconciliation Architecture

Apache EMR Cluster and Rule Engine:

● Apache EMR cluster utilizes the rule engine to match transactions.

● Successful reconciliation results are written back to AWS S3.

● Exceptions are sent to AWS EventBridge, which triggers a Lambda function to enrich and report exceptions back to S3.

Storage and Operational Aspects:

● Data stored as parquet files in S3.

● Apache Glue Catalog configured on top of parquet files.

● The operational team can query data using Amazon Athena in SQL fashion.

● Custom-built UI portal on top of Glue Catalog provides detailed reconciliation states and outcomes for specific days, settlements, or partners.

Architecture Highlights:

Active-Active Architecture:

● Operates across multiple AWS regions.

● Ensures high availability with zero recovery point objective (RPO) and recovery time objective (RTO).

● If one region goes down, another can process transactions seamlessly.

● In-flight transactions are managed using Amazon Kinesis Data Stream and DynamoDB for consistency across regions.

Technology and Architecture Decisions:

● AWS EMR vs. Redshift:

● Considered using Redshift for a data lake solution but opted for AWS EMR due to cost efficiency.

● EMR cluster extension to the core processing system, leveraging existing S3 data store.

● Low-cost solution achieved by using AWS EMR to realize the problem statement.


Business Impact of the New Reconciliation Solution

Accuracy:

● Improved from three 9s to four 9s.

Speed:

● Reduced reconciliation time from 24 hours to a 15-minute cycle.

● Horizontal cluster ensures consistent processing time (max 30 minutes) regardless of transaction volume (1 million to 300 million transactions).

Risk Reduction:

● Faster reconciliation (within 15 minutes) minimizes potential fraud and risk.

● Allows for quicker action on system or external issues.

Cost Optimization:

● Chosen AWS EMR cluster over data lake solutions for cost efficiency.

● EMR cluster and Lambda functions operate on-demand, not continuously.

● Computing instances have a limited lifetime, freeing up resources and minimizing costs once processing is complete.