Skip to content

VEOXVEOX | Quick Start

Everything you need to build, test, secure, and evolve with VEOX.

Table of Contents

  1. Overview
  2. Data Protection & Privacy
  3. Server Quick Start
  4. Python SDK Quick Start
  5. Per-Algorithm Guides

Overview

The VEOX system is composed of two parts:

  1. The VeoxEvolver python client -- a PyPi package you can use in your Python scripts and notebooks
  2. The server -- the "backend" engine VeoxEvolver connects to, in order to perform algorithmic discovery

This Quickstart guide will help you get both components up and running quickly.

VEOX Evolution Dashboard
Live evolution dashboard — fitness scatter, champion trend, and pipeline leaderboard.

Data Protection & Privacy

VEOX is designed to protect the privacy of your data and the integrity of your intellectual property (i.e. novel algorithms). This guide will help you set up a VEOX server locally on your own machine, ensuring that your data never leaves your system. All data "uploaded" to the server is merely copied to the server running on your machine. It is never shared with VEOX or any third party.

1. VEOX Server Setup

Pre-requisites: a container runtime such as Docker or Podman must be present on your system. Refer to the Docker documentation for installation instructions.

Step 1: Pull the Docker container

docker pull 714044927654.dkr.ecr.us-east-2.amazonaws.com/doug/single_enclave/veox-enclave-server:latest

Step 2: Run the Docker container

docker run -d --name veox-enclave-server -p 8090:8090 714044927654.dkr.ecr.us-east-2.amazonaws.com/doug/single_enclave/veox-enclave-server:latest

Create a docker-compose.yml file:

version: '3.8'
services:
  veox-enclave-server:
    image: 714044927654.dkr.ecr.us-east-2.amazonaws.com/doug/single_enclave/veox-enclave-server:latest
    ports:
      - "8090:8090"

Then run: 

docker compose up -d

Step 3: Verify It's Running

# Health check
curl http://127.0.0.1:8090/health
# → {"status": "healthy"}

from veox import VeoxEvolver

ev = VeoxEvolver("binary", api_url="http://127.0.0.1:8090")
ev.health_check()   # ✅ Server is healthy

2. Python SDK Quick Start

The VeoxEvolver client is a Python PyPi package. Dependencies: requests, rich, pandas, scikit-learn.

Step 1: Install the Python SDK

You can install it directly via pip:

pip install veox

Step 2: Import VeoxEvolver

from veox import VeoxEvolver

Step 3: Evolve with Your Own Data

from sklearn.datasets import make_classification
import pandas as pd
from veox import VeoxEvolver

# Generate a dataset (or load your own CSV)
X, y = make_classification(n_samples=500, n_features=20, n_informative=10, random_state=42)
df = pd.DataFrame(X, columns=[f"f{i}" for i in range(20)])
df["target"] = y

# Evolve! Data is uploaded to the server (ephemeral /tmp/ in Docker)
evolver = VeoxEvolver("binary")
evolver.fit(data=df, target_column="target", max_generations=5)
print(f"Best AUC: {evolver.best_fitness_:.4f}")

Key Features

Feature Code
Upload data evolver.fit(data=df, target_column="target")
Health check evolver.health_check()
List families VeoxEvolver.list_algorithms()
Custom fitness evolver.fit(custom_fitness_code="...")
Callbacks evolver.fit(on_champion=my_handler)
Export JSON evolver.to_json()
Export DataFrame evolver.to_dataframe()
Save / Load evolver.save("r.json") / VeoxEvolver.load("r.json")
Auto server with VeoxEvolver.serve() as ev: (connects to running enclave)
Silent mode VeoxEvolver("binary", verbose=False)

Result Attributes

Attribute Description
evolver.best_fitness_ Best score
evolver.best_pipeline_ Decoded pipeline: XGBoost → Scaler → Fusion
evolver.champions_ All champions with scores
evolver.result_ Full EvolutionResult container
evolver.result_.evals_per_minute Throughput
evolver.result_.success True if completed with ≥ 1 eval

3. Per-Algorithm Guides

Detailed guides for each algorithm family with pipeline stages, fitness configuration, baselines, and tips:

Family Guide Metric
🎯 Binary Classification Binary Classification AUC (Area Under ROC Curve)
📈 Regression Regression R² (coefficient of determination)
⚡ Optimization Optimization Best function value (lower = better)
🔍 Outlier Detection Outlier Detection PR-AUC (Area Under Precision-Recall Curve)
⏳ Time Series Time Series MASE
📊 Trading Trading Sharpe Ratio
🔊 Signal Separation Signal Separation SINR (dB)
---
Guide Description
Comparison Multi-family Comparison
Save & Load Data Save & Load Result Data
Server Connection Automatic & Reliable Server Connection
Custom Fitness Custom Fitness Functions
Paper API Automate paper generation — PDF is built inside the enclave via API; no LaTeX on your machine

Testing (verified)

All 7 algorithm families (binary, regression, optimization, outlier, trading, time_series, signal_separation) are verified against a running enclave via the E2E API smoke test. Free tier (no activation) is sufficient for Public Examples and the smoke test. PRO tier (license activation) is required for PaperKit/PDF and ProExamples.

What Command / doc
Start enclave bash Private/scripts/run_enclave.sh
Health check curl -s http://127.0.0.1:8090/health
Full API smoke (Docker) See E2E Evolution Test for the exact docker run command