Automated Trading System

2023 – present

Real-money algorithmic trading platform for US100/NASDAQ-100 CFDs via Capital.com, evolved from a data pipeline into a full agent framework with a charting dashboard.

600+ ticks/min processed in real-time

3 independent services running in production

Sub-second dashboard SSR response times

The Challenge

Needed reliable real-time data ingestion, backtesting infrastructure with realistic fee modeling, and a modern dashboard, all self-hosted and maintained independently.

Approach

Built a TypeScript/Bun backend handling real-time WebSocket streams and batch REST backfills into PostgreSQL. Designed a modular agent framework where trading strategies are drop-in files with automated backtesting and a leaderboard. Frontend is a SvelteKit dashboard with SSR and a standalone PWA charting app optimized for iOS.

Key Highlights

Real-time WebSocket tick capture with dual-buffer async batching
Synthetic tick generation from OHLC candles for backtesting at 100ms resolution
Drop-in strategy framework with automated backtesting and leaderboard
SvelteKit dashboard with equity curves and agent management
iOS PWA with custom app state engine (iOS 18 + 26 compatible)

Technologies

TypeScriptBunSvelteKit 2Svelte 5PostgreSQLWebSockets