Overview
The rise of cryptocurrencies and their growing market capitalization has prompted many traditional investors to analyze these assets quantitatively and include them in their portfolios. However, the rapid emergence of new cryptocurrencies and exchanges, combined with their inherent volatility, makes trading these assets complex and often inaccessible to many investors.
Lacking in-house expertise to build a robust platform for backtesting crypto trading strategies, our client, a Tier 1 Banking group engaged DataSpartan. We led the development of their first in-house crypto trading platform, capable of operating across multiple exchanges and supporting market-making activities in this rapidly evolving space.
Approach
We assembled a multidisciplinary team of software developers and traders to build a platform tailored to the unique demands and constraints of the crypto market. Beyond software development, the team analyzed both intrinsic and technical risks and created the initial prototypes of professional trading strategies.
The backtesting system was implemented in Java, leveraging Big Data technologies such as Spark Streaming, Kafka, and HDFS to support highly demanding exchange ETL pipelines. A continuous data collection framework was established, sourcing information from exchanges and other relevant channels, including social media, resulting in proprietary datasets and trading signals. The live trading infrastructure was developed using C++ and other low-latency technologies to ensure optimal performance.
Results
The system is fully operational and continues to support our quantitative team in researching profitable strategies for both high- and medium-frequency market-making. Several market-tested strategies with above-market Sharpe ratios have already been identified and are actively traded by our partners.
