Back to Projects
Flagship case studyreview surface

Meridian

A premium trading signal platform combining Convex for real-time reactivity, TimescaleDB for time-series analytics, and Interactive Brokers integration for trade execution. Meridian features multi-source data ingestion, human-in-the-loop approval workflows, and a live dashboard for monitoring signal performance.

Trading PlatformReal-Time DataConvexTimescaleDBIBKR
Meridian

Build surface

The implementation surface for this system. These are the layers that mattered in practice, not a generic skills wall.
Real-time
Convex, WebSocket
Database
TimescaleDB, time-series data
Trading
IBKR integration, market data
Analytics
Signal processing, approval workflows
Frontend
Next.js, React, TailwindCSS

System Architecture

Meridian's architecture flows from multi-source data ingestion through signal processing and approval workflows to trade execution via IBKR, with Convex providing real-time reactivity for the analytics dashboard.

Trading Signal Pipeline

Market data from multiple sources flows through signal processing and a human-in-the-loop approval workflow before trade execution via IBKR. TimescaleDB stores time-series data while Convex pushes real-time updates to the dashboard via WebSocket.

Why This Stack?

Convex provides real-time reactivity without manual WebSocket management, making dashboard updates instant. TimescaleDB handles time-series data with continuous aggregates for historical analytics. IBKR's TWS API offers institutional-grade market data and execution capabilities. Together, they form a platform that is both responsive and analytically powerful.

Core Features

01

Multi-Source Data Ingestion

Aggregate trading signals from multiple data sources into a unified processing pipeline

  • Real-time market data feeds via IBKR TWS API
  • Alternative data source integration for signal diversity
  • Normalized data pipeline with schema validation
  • Historical data backfill from TimescaleDB for backtesting
02

Approval Workflows

Human-in-the-loop approval system ensuring every trade signal is reviewed before execution

  • Multi-stage approval pipeline with configurable reviewers
  • Signal confidence scoring to prioritize review queue
  • Time-sensitive expiration for signals with market relevance windows
  • Audit trail of all approval decisions with reasoning capture
03

Real-Time Analytics

Live dashboard powered by Convex and WebSocket for instant visibility into signal performance

  • Convex reactive queries for real-time dashboard updates
  • WebSocket push for real-time P&L and position tracking
  • TimescaleDB continuous aggregates for historical analytics
  • Customizable chart views with TailwindCSS-styled components
04

IBKR Integration

Deep integration with Interactive Brokers for market data consumption and trade execution

  • TWS API connection for real-time market data streaming
  • Order management with support for multiple order types
  • Position tracking and portfolio synchronization
  • Account data monitoring for risk management

Platform Metrics

Signals
Real-time
WebSocket-driven updates
Data Ingestion
Multi-source
Diverse signal origins
Workflows
Approval
Human-in-the-loop
Integrated
IBKR
Trade execution ready

Human-in-the-Loop Design

Meridian's approval workflow ensures no trade executes without human review. Signals are scored by confidence, queued for review with market-relevant expiration windows, and logged with full audit trails. This design balances automation speed with the judgment that algorithmic systems alone cannot provide.

Time-Series Analytics

TimescaleDB's hypertables and continuous aggregates enable efficient querying across time-series data. Combined with Convex's reactive subscriptions, the dashboard delivers both historical depth and real-time responsiveness for informed trading decisions.

Next inspection step

Inspect the system further

Use the live surface or the source as the next level of proof. The goal here is not to end on a marketing flourish, but to make the next inspection step obvious.

Source
https://github.com/hopeatina
Why this matters
Strong systems work should be inspectable from multiple angles: interface, architecture, and implementation.