Citadel Sector Pods AI Trading Team

This strategy is inspired by the pod-style structure associated with Ken Griffin’s Citadel and other multi-manager platforms. The idea is simple: do not ask one generalist to understand every market at once. Give each specialist a clear lane, let them pitch their strongest idea, then put a risk manager and portfolio manager above the debate.

In Lumibot, that becomes an AI trading team. Sector pods study different parts of the market, a risk manager challenges the strongest pitches, and only the portfolio manager has permission to submit orders. It is a good example when you want to test whether specialist agents can create better decisions than one single broad prompt.

See this strategy running live on BotSpot

How the team works

• technology_pod looks at technology and communications exposure.

• financials_pod looks at financials and rate-sensitive exposure.

• healthcare_pod looks at healthcare and defensive growth.

• energy_pod looks at energy and commodity-sensitive exposure.

• consumer_pod looks at consumer and housing-sensitive exposure.

• risk_manager challenges crowding, drawdown, macro, and reversal risk.

• portfolio_manager rotates into the strongest sector ETF and is the only agent allowed to trade.

Backtest snapshot

Run it with a broker

The file defaults to broker-connected execution. With Alpaca, it runs in paper mode unless you set ALPACA_IS_PAPER=false.

export GEMINI_API_KEY='your-key-here'
export ALPACA_API_KEY='your-alpaca-key'
export ALPACA_API_SECRET='your-alpaca-secret'
export ALPACA_IS_PAPER=true
python lumibot/example_strategies/ai_trading_team_citadel_sector_pods.py

Backtest it

Use the same strategy class and change IS_BACKTESTING = False to IS_BACKTESTING = True in the runner:

export GEMINI_API_KEY='your-key-here'
python lumibot/example_strategies/ai_trading_team_citadel_sector_pods.py

Example code

"""Citadel / Surveyor-inspired sector-pod AI trading team example.

This example is inspired by public descriptions of sector-specialist equities
teams. It is not affiliated with or endorsed by Citadel, Surveyor Capital, or
related companies.

Set GEMINI_API_KEY plus Alpaca credentials, then run paper trading:
    python ai_trading_team_citadel_sector_pods.py

Set IS_BACKTESTING=True in the runner to run the historical example instead.
"""

import os
from datetime import datetime

from lumibot.strategies.strategy import Strategy


class AITradingTeamCitadelSectorPodsStrategy(Strategy):
    parameters = {
        "universe": ["XLK", "XLF", "XLV", "XLE", "XLY", "XLI", "XLP", "XLU", "XLB", "XLRE", "XLC"],
    }

    def initialize(self):
        self.sleeptime = "1D"
        model = os.environ.get("AI_TRADING_TEAM_MODEL", "gemini-3.1-flash-lite")
        self.agents.create(
            name="technology_pod",
            model=model,
            allow_trading=False,
            system_prompt="Rank technology and communications sector ETFs. Use exact symbols from the universe only. Do not invent symbols or table names.",
        )
        self.agents.create(
            name="financials_pod",
            model=model,
            allow_trading=False,
            system_prompt="Rank financial and rate-sensitive sector ETFs. Use exact symbols from the universe only. Do not invent symbols or table names.",
        )
        self.agents.create(
            name="healthcare_pod",
            model=model,
            allow_trading=False,
            system_prompt="Rank healthcare and defensive growth sector ETFs. Use exact symbols from the universe only. Do not invent symbols or table names.",
        )
        self.agents.create(
            name="energy_pod",
            model=model,
            allow_trading=False,
            system_prompt="Rank energy and commodity-sensitive sector ETFs. Use exact symbols from the universe only. Do not invent symbols or table names.",
        )
        self.agents.create(
            name="consumer_pod",
            model=model,
            allow_trading=False,
            system_prompt="Rank consumer discretionary, staples, and housing-sensitive sector ETFs. Use exact symbols from the universe only. Do not invent symbols or table names.",
        )
        self.agents.create(
            name="risk_manager",
            model=model,
            allow_trading=False,
            system_prompt="Compare the pod picks. Challenge crowding, factor exposure, drawdown risk, and reversal risk.",
        )
        self.agents.create(
            name="portfolio_manager",
            model=model,
            allow_trading=True,
            system_prompt="Choose the best sector ETF from the universe. Use nearly all cash and rotate decisively.",
        )

    def on_trading_iteration(self):
        context = {
            "date": self.get_datetime().date().isoformat(),
            "universe": self.parameters["universe"],
        }
        technology = self.agents["technology_pod"].run(task_prompt="Pick the best technology or communications ETF.", context=context)
        financials = self.agents["financials_pod"].run(task_prompt="Pick the best financial or rate-sensitive ETF.", context=context)
        healthcare = self.agents["healthcare_pod"].run(task_prompt="Pick the best healthcare or defensive-growth ETF.", context=context)
        energy = self.agents["energy_pod"].run(task_prompt="Pick the best energy or commodity-sensitive ETF.", context=context)
        consumer = self.agents["consumer_pod"].run(task_prompt="Pick the best consumer or housing-sensitive ETF.", context=context)
        risk = self.agents["risk_manager"].run(
            task_prompt="Compare all pod picks and identify the biggest risks.",
            context={**context, "technology": technology.summary, "financials": financials.summary, "healthcare": healthcare.summary, "energy": energy.summary, "consumer": consumer.summary},
        )
        self.agents["portfolio_manager"].run(
            task_prompt="Sell anything that is not the strongest sector ETF, then buy the strongest ETF with nearly all available cash.",
            context={**context, "technology": technology.summary, "financials": financials.summary, "healthcare": healthcare.summary, "energy": energy.summary, "consumer": consumer.summary, "risk": risk.summary},
        )


if __name__ == "__main__":
    IS_BACKTESTING = False

    if IS_BACKTESTING:
        from lumibot.backtesting import YahooDataBacktesting

        AITradingTeamCitadelSectorPodsStrategy.backtest(
            YahooDataBacktesting,
            datetime(2026, 4, 7),
            datetime(2026, 5, 22),
        )
    else:
        from lumibot.brokers import Alpaca
        from lumibot.traders import Trader

        ALPACA_CONFIG = {
            "API_KEY": os.environ["ALPACA_API_KEY"],
            "API_SECRET": os.environ["ALPACA_API_SECRET"],
            "PAPER": os.environ.get("ALPACA_IS_PAPER", "true").lower() != "false",
        }

        broker = Alpaca(ALPACA_CONFIG)
        strategy = AITradingTeamCitadelSectorPodsStrategy(broker=broker)

        trader = Trader()
        trader.add_strategy(strategy)
        trader.run_all()