Artificial intelligence for quantitative trading

Abstract

This report presents the development and enhancement of NTU Quant AI, an AI-driven quantitative trading platform that facilitates trading across multiple asset classes, including traditional assets like stocks and emerging markets such as cryptocurrencies, within a unified interface. Driven by the rapid adoption of algorithmic trading and the increasing need for customizable trading strategies, NTU Quant AI addresses existing market limitations by allowing users to deploy personalized Python-based trading algorithms seamlessly. This project iteration focuses on developing the core foundational components of the NTU Quant AI platform. Key improvements include the implementation of a robust and scalable ETL pipeline for cryptocurrency data, front-end components for user interaction, a strategy execution pipeline, and a manual data imputation module to enable strategy testing. Significant emphasis is placed on the design and implementation of the ETL pipeline, which systematically extracts cryptocurrency data from the CryptoCompare API, transforms this data into actionable metrics such as daily averages and price changes, and loads it efficiently into TimescaleDB. Strategic design choices, such as modular scheduling through Kotlin Coroutines and the separation of raw and transformed data, facilitate scalability, reliability, and data integrity. Beyond the technical implementation, this report covers integral software engineering practices applied throughout the project, including SCRUMBAN (which combines SCRUM and KANBAN) methodologies, microservices architecture, automated testing, and effective logging and monitoring strategies. These methodologies ensure robust performance, ease of maintenance, and future scalability. Finally, the report outlines future opportunities for enhancing NTU Quant AI's capabilities, particularly through real-time data stream integration, the addition of diverse financial data sources, advanced data visualization tools, and reinforced security measures. The detailed insights gained and presented in this report offer valuable guidance for continued development and deployment, solidifying NTU Quant AI’s role as a flexible and powerful tool for modern quantitative trading.