Skip to content
Maxime Grenu edited this page Feb 18, 2026 Β· 2 revisions

ELVIS β€” Algorithmic Trading Bot

⚠️ NOT FOR LIVE TRADING β€” ELVIS is an experimental research system. All strategies run in paper trading mode only. Never deploy with real funds without fully understanding the risks. Algorithmic trading can result in total loss of capital.

Overview

ELVIS is a modular, ML-driven algorithmic trading system for cryptocurrency markets. It integrates multiple machine learning architectures and supports real-time data processing, risk management, backtesting, and paper trading simulation.

Repository: cluster2600/ELVIS


Sprint Status

Sprint Focus Status
Sprint 1 Core ML models, paper trading harness, security hardening βœ… Done
Sprint 2 Reinforcement learning (PPO/DQN), live data feeds πŸ”„ In Progress
Sprint 3 Multi-exchange support, portfolio-level risk πŸ“‹ Planned
Sprint 4 Strategy optimization, walk-forward validation πŸ“‹ Planned
Sprint 5 Production-grade monitoring, kill-switch hardening πŸ“‹ Planned

Key Metrics (Paper Trading β€” Sprint 1)

Metric Value
Sharpe Ratio 0.69
Max Drawdown 0.009%
Win Rate 70%
Mode Paper trading only

Quick Links

  • Setup β€” Installation and configuration
  • Architecture β€” System design and component diagram
  • Paper Trading β€” Results and how to run
  • Security β€” Vault, secrets, and kill-switch
  • Roadmap β€” Sprint 1–5 plan

Clone this wiki locally