Skip to content

TradAI

Quantitative trading platform built with Python, Freqtrade, ArcticDB, and AWS.

Python FastAPI Freqtrade ArcticDB AWS

Quick Start

  • Your First Backtest

    Run a strategy backtest end-to-end in under 2 minutes.

    Run backtest

  • Create a Strategy

    Scaffold and implement a custom trading strategy in 5 minutes.

    Create strategy

  • Collect Market Data

    Sync OHLCV data from exchanges into ArcticDB in 5 minutes.

    Collect data

Key Features

  • Strategy Backtesting

    Run backtests with Freqtrade, compare results, optimize hyperparameters, and validate with walk-forward analysis.

  • Production-Grade Architecture

    3-layer clean architecture with Protocol-based DI, Pydantic models, and SOLID principles throughout.

  • AWS-Native Deployment

    ECS Fargate, 18 Lambda functions, Step Functions workflows, DynamoDB, S3, and Pulumi IaC.

  • ML Lifecycle

    MLflow experiment tracking, model drift detection (PSI), automated retraining, and champion/challenger comparison.

  • High-Performance Storage

    ArcticDB on S3 for versioned time-series data with incremental sync and coverage tracking.

  • Comprehensive Testing

    15 test suites, 60% coverage enforced (80% target), financial accuracy tests, and load testing with Locust.

Architecture at a Glance

graph LR
    CLI[CLI] --> Backend[Backend API]
    Backend --> Strategy[Strategy Service]
    Backend --> DataCol[Data Collection]
    Strategy --> Freqtrade[Freqtrade]
    Strategy --> MLflow[MLflow]
    DataCol --> ArcticDB[ArcticDB / S3]
    DataCol --> CCXT[CCXT / Exchanges]
    Freqtrade --> ArcticDB
    MLflow --> S3_MLflow[S3 / Artifacts]
    Backend --> DynamoDB[DynamoDB]

Full architecture overview

New to TradAI?

Start here

Follow the New Developer Onboarding guide to go from zero to productive in about 2 hours. It walks you through environment setup, your first backtest, strategy creation, and understanding the architecture.

Documentation Map

  • Developer Guides

    Step-by-step development workflows: setup, strategy creation, backtesting, validation, deployment.

    Guides

  • Libraries & Services

    Python libraries (tradai-common, tradai-data, tradai-strategy) and microservice documentation.

    Libraries | Services

  • Architecture

    System design, VPC networking, security, Step Functions, cost analysis, and 20+ architecture docs.

    Architecture

  • API Reference

    REST API endpoints, auto-generated SDK reference, CLI commands, and environment variables.

    Reference

  • Operations

    Runbooks for incident response, troubleshooting guides, and operational procedures.

    Runbooks

  • Technical Design

    Detailed DESIGN.md documents for every library and service with architecture decisions.

    Design

Essential Commands

just setup       # Install Python 3.11, deps, pre-commit hooks
just check       # Run lint + typecheck + tests (run before every commit)
just test        # Run all tests in parallel
just up          # Start Docker Compose services
just down        # Stop Docker Compose services
just doctor      # Verify development environment health

All commands use just

See the full command reference or run just --list to see every available recipe.