Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Orchestration Framework for Financial Agents: From Algorithmic Trading to Agentic Trading

Jifeng Li, Arnav Grover, Abraham Alpuerto, Yupeng Cao, Xiao-Yang Liu

TL;DR

This work reframes traditional algorithmic trading as an end-to-end FinAgent orchestration problem, mapping AT pipeline components to autonomous agents and introducing a memory module for auditability. It defines MCP for orchestrator–agent control and A2A for inter-agent communication, enabling a transparent, replayable workflow. The framework uses modular agent pools (data, alpha, risk, portfolio, execution, backtesting, memory) and a plan graph to run walk-forward backtests while preventing leakage to LLMs. Empirical demonstrations on stock and BTC tasks show favorable risk-adjusted performance relative to baselines and clear pathways for longer horizons and broader markets; code is available on GitHub to support replication and extension.

Abstract

The financial market is a mission-critical playground for AI agents due to its temporal dynamics and low signal-to-noise ratio. Building an effective algorithmic trading system may require a professional team to develop and test over the years. In this paper, we propose an orchestration framework for financial agents, which aims to democratize financial intelligence to the general public. We map each component of the traditional algorithmic trading system to agents, including planner, orchestrator, alpha agents, risk agents, portfolio agents, backtest agents, execution agents, audit agents, and memory agent. We present two in-house trading examples. For the stock trading task (hourly data from 04/2024 to 12/2024), our approach achieved a return of $20.42\%$, a Sharpe ratio of 2.63, and a maximum drawdown of $-3.59\%$, while the S&P 500 index yielded a return of $15.97\%$. For the BTC trading task (minute data from 27/07/2025 to 13/08/2025), our approach achieved a return of $8.39\%$, a Sharpe ratio of $0.38$, and a maximum drawdown of $-2.80\%$, whereas the BTC price increased by $3.80\%$. Our code is available on \href{https://github.com/Open-Finance-Lab/AgenticTrading}{GitHub}.

Orchestration Framework for Financial Agents: From Algorithmic Trading to Agentic Trading

TL;DR

This work reframes traditional algorithmic trading as an end-to-end FinAgent orchestration problem, mapping AT pipeline components to autonomous agents and introducing a memory module for auditability. It defines MCP for orchestrator–agent control and A2A for inter-agent communication, enabling a transparent, replayable workflow. The framework uses modular agent pools (data, alpha, risk, portfolio, execution, backtesting, memory) and a plan graph to run walk-forward backtests while preventing leakage to LLMs. Empirical demonstrations on stock and BTC tasks show favorable risk-adjusted performance relative to baselines and clear pathways for longer horizons and broader markets; code is available on GitHub to support replication and extension.

Abstract

The financial market is a mission-critical playground for AI agents due to its temporal dynamics and low signal-to-noise ratio. Building an effective algorithmic trading system may require a professional team to develop and test over the years. In this paper, we propose an orchestration framework for financial agents, which aims to democratize financial intelligence to the general public. We map each component of the traditional algorithmic trading system to agents, including planner, orchestrator, alpha agents, risk agents, portfolio agents, backtest agents, execution agents, audit agents, and memory agent. We present two in-house trading examples. For the stock trading task (hourly data from 04/2024 to 12/2024), our approach achieved a return of , a Sharpe ratio of 2.63, and a maximum drawdown of , while the S&P 500 index yielded a return of . For the BTC trading task (minute data from 27/07/2025 to 13/08/2025), our approach achieved a return of , a Sharpe ratio of , and a maximum drawdown of , whereas the BTC price increased by . Our code is available on \href{https://github.com/Open-Finance-Lab/AgenticTrading}{GitHub}.

Paper Structure

This paper contains 20 sections, 4 equations, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Proposed Framework for FinAgent Orchestration
  3. Overview
  4. Control Messages and Agents' Communications
  5. Homegrown Trading Examples
  6. Stock Trading and Crypto Trading Tasks
  7. Backtesting Performance
  8. Conclusion
  9. More Test Details
  10. BTC/USDT Trading Strategy
  11. Feature Engineering and Data Preprocessing
  12. Model Architecture and Training Methodology
  13. Signal Generation and Market Regime Identification
  14. Position Sizing and Risk Management
  15. Backtesting Results and Performance Evaluation
  16. ...and 5 more sections

Figures (3)

  • Figure 1: Agentic trading vs. algorithmic trading: we map the AT components to agents in our FinAgent orchestration framework, where a memory agent provides the contexts to other agents.
  • Figure 2: Seven-stock cumulative returns with the test window from 24/04/2024 to 31/12/2024 (within the 2022--2024 sample, the scrolling training window size is 3 months). The agentic strategy shows lower volatility and a smaller max drawdown, while the equally-weighted benchmark attains the highest total return. ETF baselines: SPY, QQQ, IWM, VTI. Metrics are reported in Table \ref{['tab:bench_comp_ours']}.
  • Figure 3: BTC results (07/27 to 08/13 in 2025, the scrolling window is 7 days). Cumulative returns: Buy&Hold $+3.80\%$, Ours $+8.39\%$, Excess $+4.59\%$. Excess $=$ Ours $-$ Buy&Hold.