case study

crypto market maker simulator

A full-stack crypto market-making simulator for BTC/USDT and ETH/USDT, with configurable two-sided quotes, inventory-aware strategy, simulated fills, portfolio accounting, and real-time market monitoring.

overview

An educational market-making desk simulation built end to end. A configurable quote engine places simulated bids and asks around the market price, accounts for inventory skew and hard risk limits, and tracks fees, positions, and realised and unrealised P&L. The deterministic default feed makes sessions reproducible, while an opt-in Kraken public ticker provides live prices without placing real orders.

architecture

React / Vite / TypeScript / Tailwind dashboard → Nginx → FastAPI REST and bounded WebSocket updates → async SQLAlchemy / Alembic → PostgreSQL. Market data is provided by a deterministic simulator by default or Kraken's public v2 ticker when enabled.

engineering focus

Applying Decimal arithmetic to financial quantities and keeping accounting values precise across API boundaries.
Designing inventory-aware quoting and explicit risk controls that suppress quotes when market data is stale or limits are breached.
Separating market-data providers behind an interface so simulated and public live feeds share the same engine.
Coordinating async REST APIs, bounded WebSocket events, persistence, migrations, and a real-time React dashboard.
Making the system reproducible and testable through deterministic simulation and CI.

highlights

Decimal-based quote generation with configurable spread, quote size, inventory skew, and inventory/exposure limits.
Simulated quote-crossing fills, fees, long/short positions, average entry price, and realised/unrealised P&L.
Live dashboard for market prices, bid/ask quotes, strategy versions, portfolio, fills, and risk headroom, updated over WebSockets.
Deterministic simulated market feed plus optional Kraken v2 public ticker with reconnect backoff and stale-data detection.
PostgreSQL persistence for sampled market observations, immutable strategy versions, simulated trading state, and trade history.
Docker Compose deployment and CI checks for backend and frontend linting, types, tests, and builds.

project context

Simulation and education only: no real orders or funds are used, and the public market-data provider needs no private exchange API credentials.

project media

Existing portfolio media

DISCUSS A PROJECT

overview

engineering focus

Applying Decimal arithmetic to financial quantities and keeping accounting values precise across API boundaries.

Designing inventory-aware quoting and explicit risk controls that suppress quotes when market data is stale or limits are breached.

Separating market-data providers behind an interface so simulated and public live feeds share the same engine.

Coordinating async REST APIs, bounded WebSocket events, persistence, migrations, and a real-time React dashboard.

Making the system reproducible and testable through deterministic simulation and CI.

highlights

Decimal-based quote generation with configurable spread, quote size, inventory skew, and inventory/exposure limits.

Simulated quote-crossing fills, fees, long/short positions, average entry price, and realised/unrealised P&L.

Live dashboard for market prices, bid/ask quotes, strategy versions, portfolio, fills, and risk headroom, updated over WebSockets.

Deterministic simulated market feed plus optional Kraken v2 public ticker with reconnect backoff and stale-data detection.

PostgreSQL persistence for sampled market observations, immutable strategy versions, simulated trading state, and trade history.

Docker Compose deployment and CI checks for backend and frontend linting, types, tests, and builds.