Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/najmulhossainnj/Hedge-fund-backend/llms.txt

Use this file to discover all available pages before exploring further.

The Hedge Fund Backend is a production-grade Python service that powers a full quant research workflow. From defining trading strategies and engineering features to training models, running backtests, and promoting validated strategies to a portfolio layer — every step is accessible via a clean REST API at /api/v1.

Quickstart

Run the service locally and make your first API calls in minutes.

Architecture

Understand how the engines, plugins, workers, and event bus fit together.

API Reference

Browse every endpoint — strategies, features, models, backtests, and more.

Plugin System

Add custom features, models, signal generators, and backtest engines.

What’s Inside

The platform is organized into distinct research layers, each exposed through its own API group:

Strategies

Define, version, and promote trading strategies through their full lifecycle.

Feature Engine

Generate versioned feature datasets from OHLCV and news data with content-hash deduplication.

Model Training

Train XGBoost, LightGBM, CatBoost, LSTM, and Random Forest models with time-series CV and Optuna tuning.

Signal Generation

Convert model predictions into BUY/SELL/HOLD signals using rule trees or plugin generators.

Backtesting

Run backtests with vectorbt or Backtrader, computing 14+ performance, risk, and trading metrics.

Validation

Gate strategies through walk-forward analysis and CPCV with PBO and Deflated Sharpe checks.

Getting Started

1

Start the infrastructure

Spin up PostgreSQL, Redis, MinIO, and MLflow using Docker, then run Alembic migrations.
2

Launch the API server

Start the FastAPI service with uvicorn app.main:app --reload and visit /docs for the interactive Swagger UI.
3

Create your first strategy

POST to /api/v1/strategies to define a strategy, then attach features, train a model, and run a backtest.
4

Run the AI Research Agent

Send a natural language query to POST /api/v1/agents/research and let the agent orchestrate the entire pipeline automatically.
The service exposes interactive API documentation at http://localhost:8000/docs (Swagger UI) and http://localhost:8000/redoc (ReDoc) when running locally.

Build docs developers (and LLMs) love