Automated Trading Agent
From a Jupyter Notebook to a live cloud system – this project genuinely changed how I think about data science.
For my MSc project, I wanted to build something that did not just sit in a static file. I wanted it to run on its own, every single day, without manual interference. So I built a fully automated trading dashboard on AWS, and honestly, putting it together taught me more than hours of coursework.
Here is how it works every morning at 8:00 A.M. Local Time:
⏰ AWS EventBridge triggers the pipeline automatically at 8:00 A.M. Local Time.
📈 Live open, high, low, close, and volume market data is fetched alongside global financial news.
🧠 NLP models score news sentiment, while a Random Forest evaluates technical signals across all 11 assets.
🛡️ A risk agent filters everything and logs approved trades to the AWS RDS database.
📊 The Streamlit dashboard on EC2 refreshes automatically with the day's signals.
The most valuable part? Figuring out AWS Security Groups, VPC configurations, and how to get all these services to actually communicate with each other.
This project reminded me that there is a significant gap between training a model and deploying one. Bridging that gap is where the real learning happens.
If you are working on something similar or have gone through the same cloud architecture headaches, I would genuinely love to hear about your experience. 👇
Comments
Post a Comment