Laith Al Enooz
← All work

INGOT Brokers · 2021–present

High-Frequency Trading Platform

A core trading platform built from scratch for 100K+ concurrent users, with ultra-low-latency execution and a full audit trail.

100K+ concurrent users · sub-100ms inter-service latency

  • PHP
  • Go
  • PostgreSQL
  • Redis
  • Kafka
  • AWS
orders fills events Clients 100K+ concurrent Order Execution Position Mgmt Kafka bus Risk Redis cache / session
Follow it through the system, step by step an interactive teardown — scroll, and the architecture assembles →

The problem

A brokerage needs a trading core that never loses an order, never double-fills, and never lies about a position, under load that spikes with the market rather than the deploy schedule. The existing stack couldn’t hold the concurrency or give operations a trustworthy audit trail.

What I built

I architected the platform from the ground up: order execution and position management designed to hold 100K+ concurrent users with ultra-low-latency execution and real-time position tracking.

  • Event-driven core. Trading, risk, and CRM services communicate over Kafka, decoupling the hot execution path from everything downstream and keeping inter-service latency under 100ms.
  • Redis for the session and caching layer, keeping hot reads off the primary database.
  • A full audit trail of every order state transition, non-negotiable for a regulated financial system and the thing that lets operations trust the numbers.
  • Database performance work across PostgreSQL and the wider data layer: indexing strategy, partitioning, and query optimization tuned for transaction throughput.

The result

The execution patterns I established became the reference design adopted across multiple backend teams. The platform handles production peak load with headroom, and the audit trail has held up to the scrutiny a financial system attracts.