Order Management System (OMS)

A long-running flagship project: a modular, event-driven Order Management System designed for capital markets. It simulates the full trade lifecycle — from order intake and risk checks to matching, trade generation, and reporting — while integrating real-time market data streams.

Goals and Motivation

The motivation behind this project was to deepen my expertise in backend architecture for trading systems and to build a realistic, production-grade platform that demonstrates:

• Event-driven design principles
• Modular service boundaries
• Real-time data integration
• Extensibility for future enhancements (e.g., Rust modules, Kubernetes scaling)

It also serves as a portfolio anchor, showcasing my ability to design, implement, and document complex distributed systems with recruiter-ready clarity.

System Architecture

The OMS is composed of five core services, each containerized and orchestrated with Docker Compose:

• Order Service
  Handles order intake via REST (/api/orders), validates payloads, and publishes to the orders Kafka topic.

• Risk Service
  Consumes both orders and quotes topics. Performs pre-trade risk checks (e.g., price bands, quantity limits) using the latest market data before forwarding valid orders.

• Matching Engine
  Maintains an in-memory order book, consumes validated orders, and generates trades by matching buy/sell orders. Publishes trades to the trades topic.

• Core Service
  Persists orders and trades to storage, provides reporting endpoints, and simulates downstream integration with clearing/settlement systems.

• Market Data Service
  Streams synthetic or replayed quotes into the quotes topic. Provides optional REST snapshots (/api/quotes) for debugging and validation.

Kafka Topics and Schemas

• quotes → symbol, bid/ask, last price, timestamp
• orders → orderId, symbol, side, quantity, price, tif, timestamp
• trades → tradeId, orderIds, symbol, quantity, price, timestamp

All messages use Avro schemas with schema evolution support for forward/backward compatibility.

Key Features

• 🧩 Microservice modularity: Clear separation of concerns across order, risk, matching, core, and market data services.
• 🔄 Event-driven architecture: Kafka as the backbone for asynchronous, decoupled communication.
• 🐳 Containerized deployment: Docker Compose orchestrates infra (Kafka, Schema Registry, Grafana) and services with shared networks and startup sequencing.
• 📡 Market data integration: Real-time quotes drive risk checks and matching logic, simulating real-world trading conditions.
• 🛡️ Risk validation: Pre-trade checks ensure only valid orders reach the matching engine.
• ⚡ Matching engine: Implements price-time priority matching with trade generation.
• 🗄️ Persistence and reporting: Core service stores trades and provides reporting APIs.
• 📖 Documentation-first approach: Architecture diagrams, setup instructions, and benchmarking workflows included.
• 🛠️ Clean design principles: DTO layering, MapStruct mapping, and separation of concerns for maintainability.

Development and Setup

• Languages & Frameworks: Java 21, Spring Boot, Maven multi-module project.
• Messaging: Kafka with Avro schemas and Schema Registry.
• Infrastructure: Docker Compose for infra + services, with environment-specific application-docker.properties.
• Monitoring: Grafana dashboards for JVM metrics, Kafka lag, and throughput.
• Testing & Benchmarking: JMeter for load generation; synthetic quote generator for market data.

Lessons Learned

• Event-driven design: Learned to balance modularity with reliability in distributed systems.
• Schema evolution: Gained practical experience with Avro and the importance of compatibility in cross-service communication.
• Real-time data pipelines: Understood how market data feeds drive risk and matching logic in trading systems.
• Infrastructure orchestration: Containerizing and sequencing multi-service environments taught me resilience and debugging across OSes.
• Documentation as a skill: Writing recruiter-ready docs and diagrams sharpened my ability to present complex systems clearly.