Skip to content
1 The short version

I build systems that move money and data, and I don't trust them until I've measured them.

I started out building WordPress sites and freelancing, then spent the last years in fintech: payment links, ledgers and settlement engines where "almost right" is still wrong. Today I bring the same habit to data platforms and the AI agents that sit on top of them.

2 Work

Senior Software Engineer (Backend)

Funavry Technologies · now

  • A unified transaction ledger across deposits, withdrawals and swaps, with an audit trail that can't be edited
  • Single-use payment links with atomic consumption and multi-provider failover
99.9% uptime · swap failures cut 95%

Senior Backend Engineer (contract)

Normies · crypto and traditional finance in one app

  • A settlement engine with atomic guarantees across Solana, Ethereum and Polygon
  • KYC and AML infrastructure, and 24/7 observability
query latency down 60% with Redis

Senior Software Engineer

NayaPay · payments for Pakistani merchants

  • Merchant onboarding and payments APIs, built around a transaction state machine
  • QR-compatible payment links with idempotency and recovery
sub-100 ms transitions · errors 12% → 2%

Before fintech

Freelance web work and technical SEO

  • Hundreds of WordPress builds and fixes on Envato Studio
  • Technical SEO for Super Stars Bio
18% revenue lift from SEO
3 How I work
Measure itA number beats an opinion. I time it, count it, and write down what broke.
Every row accounted forIn payments and data, a silent drop is the worst bug. I design for none.
Small, boring partsIdempotent steps, clear states, plain SQL. Easy to reason about at 3 a.m.
Guardrails for agentsAI agents get real tools, and real brakes on anything that costs money.
4 Stack
PythonTypeScriptNode.jsNestJSPostgreSQLRedisKafkaApache IcebergTrinoDatabricksSnowflakeSparkdbtAzureAWSDocker

// BSCS, University of Karachi