As you can see above in the architecture, this post explains how to build a serverless trading signal system using AWS Lambda, Python, alpaca‑py, Amazon S3, and Amazon EventBridge Scheduler, structured clearly into prerequisites, steps, and benefits.

Prerequisites

Accounts & Access

AWS account with permission to create:
- Lambda
- IAM roles
- S3 buckets
- EventBridge rules
Alpaca account (paper trading strongly recommended)

Local Development

Python 3.10+
Basic Python knowledge
Basic understanding of AWS services (Lambda, IAM, S3)

Python Libraries

alpaca-py
pandas
boto3

Steps

Step 1: Create an S3 Bucket

This bucket stores generated trading signal CSV files.

Example:

s3://alpaca-trading-signals-bucket/signals/AAPL-2026-01-09.csv

Step 2: Create an IAM Role for Lambda

Attach the following permissions:

AWSLambdaBasicExecutionRole (CloudWatch logs)
Permission to write to your S3 bucket

Minimal S3 policy:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": "s3:PutObject",
      "Resource": "arn:aws:s3:::alpaca-trading-signals-bucket/*"
    }
  ]
}

Step 3: Package Python Dependencies (Lambda Layer)

A SMA Cross Strategy Example

BUY → Short SMA crosses above Long SMA
SELL → Short SMA crosses below Long SMA
HOLD → No crossover

Sample Codes

import os
import boto3
import pandas as pd
from datetime import datetime

from alpaca.data.historical import StockHistoricalDataClient
from alpaca.data.requests import StockBarsRequest
from alpaca.data.timeframe import TimeFrame

AWS clients

s3 = boto3.client(“s3”)

Environment variables

ALPACA_API_KEY = os.environ[“ALPACA_API_KEY”]
ALPACA_SECRET_KEY = os.environ[“ALPACA_SECRET_KEY”]
S3_BUCKET = os.environ[“S3_BUCKET”]

Alpaca client

data_client = StockHistoricalDataClient(
ALPACA_API_KEY,
ALPACA_SECRET_KEY
)

def generate_sma_signals(
symbol: str,
short_window: int = 5,
long_window: int = 20
) -> pd.DataFrame:
“””
Fetch market data and generate SMA crossover signals.
“””

request = StockBarsRequest(
    symbol_or_symbols=symbol,
    timeframe=TimeFrame.Day,
    limit=long_window + 5
)

bars = data_client.get_stock_bars(request).df

df = bars.reset_index()
df = df[df["symbol"] == symbol]

# Calculate SMAs
df["sma_short"] = df["close"].rolling(short_window).mean()
df["sma_long"] = df["close"].rolling(long_window).mean()

# Generate signals
df["signal"] = "HOLD"

df.loc[
    (df["sma_short"] > df["sma_long"]) &
    (df["sma_short"].shift(1) <= df["sma_long"].shift(1)),
    "signal"
] = "BUY"

df.loc[
    (df["sma_short"] < df["sma_long"]) &
    (df["sma_short"].shift(1) >= df["sma_long"].shift(1)),
    "signal"
] = "SELL"

return df[[
    "timestamp",
    "close",
    "sma_short",
    "sma_long",
    "signal"
]]

def lambda_handler(event, context):
symbol = “AAPL”

signals_df = generate_sma_signals(symbol)

date_str = datetime.utcnow().strftime("%Y-%m-%d")
s3_key = f"signals/{symbol}-sma-signals-{date_str}.csv"

s3.put_object(
    Bucket=S3_BUCKET,
    Key=s3_key,
    Body=signals_df.to_csv(index=False)
)

return {
    "status": "success",
    "symbol": symbol,
    "s3_key": s3_key
}
# Upload this ZIP as a Lambda Layer and attach it to your Lambda function.

Sample output:

timestamp,close,sma_short,sma_long,signal
2026-01-05,187.22,185.91,184.88,BUY
2026-01-06,188.10,186.80,185.42,HOLD

Now we can use this csv file as a staging area to connect to a broker, however, this part will not be focus in this post.

Step 4: Schedule Execution with EventBridge

Use EventBridge Scheduler to trigger Lambda automatically.

Example (Weekdays at 9:35am ET ≈ 13:35 UTC):

cron(35 13 ? * MON-FRI *)

This makes your trading system fully automated.

Step 5: Verify

Check CloudWatch Logs for execution status
Confirm CSV files appear in S3
Validate trading signals

Benefits

Serverless & Low Cost

No servers to manage
Pay only when the function runs

Secure

API keys stored as environment variables
Least-privilege IAM permissions

Scalable & Flexible

Add more symbols easily
Extend to multiple strategies
Increase frequency without infrastructure changes

Production-Ready Pattern

Reliable scheduling with EventBridge
Durable storage with S3
Observability via CloudWatch

Easy to Extend

Execute real trades via Alpaca Trading API
Add backtesting logic
Send alerts (SNS, Slack, email)
Orchestrate workflows with Step Functions

A guide to use AWS Lambda to run a trading algorithm