← All product work

Payment Infrastructure / Case study

Cross-Border Payment Infrastructure Platform

A regulated, multi-rail payments platform processing $1B+ annual GTV and 270M+ payments a year across pay-in, payout, wallets (DCB/IBFT), card acquiring (MPGS/MDES), settlement, FX and cross-border corridors, PCI DSS and ISO/IEC 27001 certified.

$1B+Annual GTV
270M+Annual transactions
150+Merchant integrations
<0.1% GTVFraud loss

The operating decision, in context.

Owned the product, architecture and operating model of a five-market payments platform serving global enterprises and local merchants. Took a fragmented integration estate and turned it into a single regulated rail with shared APIs, ledger, settlement and risk, through a CTO departure and a regulatory tightening.

Evidence overview

What changed, and why it held.

Operator lens

Payment infrastructure is a product problem, not a platform problem, orchestration, error states and retries decide the experience.

Platform downtime

Frequent incidents→−90%

Settlement

Best-effort→Ledger-backed, corridor-aware windows

Enterprise wallet adoption

<5%→30%

−90%Downtime reduction
30%Enterprise wallet adoption
90%Straight-through processing
99.9%Uptime
◆ Diagramfig.
One regulated rail, many partners, one API contract.
RAILSUNIFIED PLATFORMCONSUMERSCardsMPGS · MDESWalletsJazzCash · EasypaisaDCBDirect carrier billingIBFT1Link · NIFTBank transfersRTGS · ACHCross-borderDLocal · Thunes · Boku · CodaUnified APIPay-in · PayoutRouting engineCost · success · healthRisk servicePre-auth · post-auth · asyncCanonical ledgerDouble-entry · idempotentSettlement engineMulti-rail · corridor-awareMerchantsConsole · WebhooksFinanceT+0 / T+1 settlementRegulatorReporting pipeline

Cards, wallets, DCB, IBFT, bank transfers and cross-border corridors converge behind a single pay-in/payout API. Routing, risk, ledger and settlement run as shared services, not per-integration code.

The job to be done.

Merchants and platforms operating in Pakistan, Bangladesh, Nepal, Sri Lanka, Iraq needed a single regulated rail to accept, payout, settle and reconcile across cards (MPGS/MDES), wallets, IBFT, DCB and cross-border corridors, without stitching together fragile point integrations. In 2024, regulatory tightening and a CTO departure forced the platform to operate under heightened scrutiny without losing pace.

What I built.

  • Unified pay-in API across cards, wallets, DCB, IBFT and bank transfers
  • Payout and disbursement engine with corridor-aware routing
  • Wallet ledger, hold/release flows and partner subaccounts
  • Cross-border + FX layer integrating DLocal, Thunes, Boku, Coda and MoneyGram
  • Settlement and reconciliation engine with T+0/T+1 reporting
  • Merchant console: onboarding, KYC/KYB, dashboards and dispute workflows

How it fits together.

  • Single pay-in/payout API surface, rails are an implementation detail behind a stable contract
  • Canonical double-entry ledger; rails post events, settlement reads from the ledger
  • Corridor abstraction with routing policy (cost, success rate, partner health)
  • Risk service in-line: pre-auth, post-auth and async monitoring share one feature store
  • Event-driven webhooks; idempotency, retries and DLQs are first-class

How it runs in practice.

  • 40-engineer payments organisation across 12 cross-functional squads
  • Weekly rail health reviews, success rate, cost, latency, dispute rate per partner
  • Joint risk + product council owning the risk taxonomy and false-positive budget
  • Regulator-facing reporting pipeline owned by product, not finance

Where I sat in the work.

Chief Product Officer (acting CTO during the 2024 regulatory tightening). Owned product strategy, roadmap, partner integrations, regulatory posture, security architecture and the org structure to ship and operate the platform end-to-end.

What moved.

  • Scaled to $1B+ annual GTV and 270M+ payments a year across 5 frontier markets
  • Onboarded enterprise platforms including TikTok, Samsung, Shein, Uber and MoneyGram
  • Held fraud loss below 0.1% of GTV; cut platform downtime by 90%
  • Led PCI DSS and ISO/IEC 27001 certification programs from scratch
  • Drove 30% enterprise wallet adoption and moved settlement onto ledger-backed, corridor-aware payout windows

What I chose against.

  • Chose a single ledger over per-rail ledgers, slower to ship rail #1, much faster after rail #3
  • Built risk in-house rather than fully outsourcing, higher ops cost, far lower false-positive rate
  • Kept enterprise-grade onboarding for low-volume merchants, accepted some activation friction to preserve regulator posture

What carries into the next build.

  • Payment infrastructure is a product problem, not a platform problem, orchestration, error states and retries decide the experience.
  • In emerging markets, local payment methods determine acceptance more than card-network features.
  • Compliance posture (PCI DSS, ISO 27001, AML/CFT) is a sales asset, not a cost line.

Relevant beyond this one platform.

The transferable core: standing up regulated multi-rail infrastructure across non-trivial markets and running it under regulator scrutiny, with partner enablement, scheme readiness, settlement, risk and reporting owned as one job at $1B+ GTV. The same job exists inside every network, PSP and cross-border platform; only the logo on the door changes.

Bring the context

Building through
real operating pressure?

Start with the decision, constraint or delivery problem in front of you. We can work out whether there is a useful fit.