Richard Mashod joins GDA Fund as the new Senior Quant Developer and Quantitative Architect Engineer.

Sydney, Australia: December 04, 2021 —

The GDA team is excited to welcome Richard as the newly assigned member of our team. His mission to build and inspire teams will amplify the production and innovation on institutional-grade crypto algorithmic trading platforms for centralized and decentralized exchanges.

Richard has been working in technology within the front office of numerous investment banks and hedge funds. Before joining GDA, he honed his skills in spectacular companies like Credit Suisse as Front Office Analyst/Developer, and with HSBC for over a year as a Market Risk Analyst/Developer. After that, he joined Bank of America Merrill Lynch as Front Office in Algorithmic Trading. He also spent seven years of his career with HFT Consultancy in Performance Engineering, which gained him a lot of expertise.

In Richard’s first few months with GDA, he has been able to innovate the following, primary and replica microservices that connect to multiple CDN to guarantee web socket feed accuracy and consistency. When the throttling policy drops the exchange connections, the consensus algorithm fills in missing data. The algorithm will also be a front-line defence in handling market manipulations, critical in an unregulated market.

Richard has been directly involved in the following developments.

We use low-level c libraries like LWS to compress the packets sent to the network; this reduces latency. On top of that, we can reduce the time it takes to react to dropped connections.

The microservice framework is highly agile and easily extendable to other exchanges. We use high-performance-computing techniques in Python and C. All our EC2 instances in our swarm/fleet have AVX512n x86 instruction set, which NumPy is optimized to use. When recreating the limit order book through snapshots and events, we vectorize the problem by using NumPy arrays and linear algebra; thus, we can keep a real-time atomic view of every event in the limit order book in memory.

We stream a large number of data features and metrics in real-time. The data features are critical in algorithmic trading and are used in reinforcement learning frameworks, especially Deep Q learning or Reinforcement Learning with a Deep Kalman Filter.

We go even further and look at order flow analytics to gauge the quality of the market. Statistically speaking, OFI is a superior predictor of mid-price change than TFI. When that relationship statically reverses, that is a good indicator of market spoofing.

Richard is very excited about his upcoming launch of a significant product with GDA:

• It improves Open Crypto Data capabilities by performance engineering and pushing the number of datasets collected by CEX, DEX & Defi ecosystems. Testing microservices and improving processes, and building adapters that handle milliseconds.
• Development of algorithmic trading strategies. On the crypto futures and perpetual.
• Development of cost-efficient adapters and data extracting solutions for smart contracts.

He’s very excited about crypto and believes in the benefits of decentralization. However, when it comes to blockchain, he thinks that although it is exciting in its infant stage, many advancements still exist.

Besides being passionate about his career, Richard does have a life outside GDA, after all. He enjoys control theory and photonic neuromorphic computing.

The GDA team is pleased to have Richard on board and can’t wait for him to achieve great things.