Selected Industry Solutions

TransQuant

High Performance Time series Computing

Distributed Computing

Multi-Model Data Research

Industry pain points

Quantitative Investment Research on big data scale

As for the traditional PC method, factors or algorithms adopted in the research could not be too complicated. The traditional method has relatively long back-testing period, inconvenience in multiple underlying assets research, long research time consumed and inconvenience in portfolio and parameter optimization.

Multi-model data quantitative investment research

With the separation among researches in technical, fundamental and informatical aspects, traditional quantitative research is based barely on trade price and volume, therefore could not dig into Alpha brought by alternative data such as public opinions, events, and satellite data. Furthermore, a single platform can't handle event-driven modeling research and investment.

High-performance distributed real-time computation

Since speed is the top priority, in the future, brokers, Quantitative PE companies or asset management institutions will all face issues to allow large-scale strategies to have distributed computing and low latency. Tick sized high performance real-time computing has narrowed strategy granularity into millisecond trading scenarios, such as real-time derivative pricing, market making, real-time portfolio management, and real-time risk management, to satisfy both low latency in trading and high throughput in data volume.

AI and Quant combined

In the past, we need a variety of complex PaaS levels such as AI platform, Quant research platform, natural language processing platform, knowledge graph platform, deep learning platform to help researchers to conduct a quantitative research on data such as public opinions.

Quantitative Investment Research on big data scale

As for the traditional PC method, factors or algorithms adopted in the research could not be too complicated. The traditional method has relatively long back-testing period, inconvenience in multiple underlying assets research, long research time consumed and inconvenience in portfolio and parameter optimization.
Multi-model data quantitative investment research

With the separation among researches in technical, fundamental and informatical aspects, traditional quantitative research is based barely on trade price and volume, therefore could not dig into Alpha brought by alternative data such as public opinions, events, and satellite data. Furthermore, a single platform can't handle event-driven modeling research and investment.
High-performance distributed real-time computation

Since speed is the top priority, in the future, brokers, Quantitative PE companies or asset management institutions will all face issues to allow large-scale strategies to have distributed computing and low latency. Tick sized high performance real-time computing has narrowed strategy granularity into millisecond trading scenarios, such as real-time derivative pricing, market making, real-time portfolio management, and real-time risk management, to satisfy both low latency in trading and high throughput in data volume.
AI and Quant combined

In the past, we need a variety of complex PaaS levels such as AI platform, Quant research platform, natural language processing platform, knowledge graph platform, deep learning platform to help researchers to conduct a quantitative research on data such as public opinions.

TRANSQUANT Solutions to Intelligent Quantitative Investment Research

Solution Highlights

High performance

time series calculation

Based on the distributed computing framework, it realizes the optimization of financial engineering algorithm based on heterogeneous hardware (CPU cluster or GPU cluster) to solve the time series calculation performance problems under large-scale data such as factor processing, quantitative backtesting, parameter optimization and combinatorial optimization.

Big date scale

distributed computing

For both classical financial engineering algorithms and derivative pricing algorithms based on new hardware framework, it has the technical capability of fully coupled heterogeneous framework. It eliminates the need for researchers to develop bottom layer database or even frameworks like GPU.

Event driven engine

bridge NLP+KG+QUANT

With a single platform, Transwarp TransQuant, an entire process from event handling, event labeling, event backtesting, strategy development, model testing to Grayscale actual quotations can be addressed.

Multi-model

data research

In the future more premium Alpha will be driven by more alternative data, such as air and outer space data including that from spacecrafts, logistics and ports, as well as space satellite data and meteorological data, etc. so that our quantitative intelligent agent can perceive investment opportunities brought by more data.

High performance time series calculation

Based on the distributed computing framework, it realizes the optimization of financial engineering algorithm based on heterogeneous hardware (CPU cluster or GPU cluster) to solve the time series calculation performance problems under large-scale data such as factor processing, quantitative backtesting, parameter optimization and combinatorial optimization.
Big date scale distributed computing

For both classical financial engineering algorithms and derivative pricing algorithms based on new hardware framework, it has the technical capability of fully coupled heterogeneous framework. It eliminates the need for researchers to develop bottom layer database or even frameworks like GPU.
Event driven engine bridge NLP+KG+QUANT

With a single platform, Transwarp TransQuant, an entire process from event handling, event labeling, event backtesting, strategy development, model testing to Grayscale actual quotations can be addressed.
Multi-model data research

In the future more premium Alpha will be driven by more alternative data, such as air and outer space data including that from spacecrafts, logistics and ports, as well as space satellite data and meteorological data, etc. so that our quantitative intelligent agent can perceive investment opportunities brought by more data.

Application Scenarios / Cases

Broker
Asset Management
Futures

Broker

Eliminating the setup of existing financial institution, for example, machine learning platform, quantitative trading platform, knowledge graph platform, deep learning platform, and multi-model data processing platform to develop AI quantitative investment research. With a single platform, TransQuant provides targeted solutions, from high performance hardware support, price-quantity quantitative research, public opinion quantitative research, to multi-model data quantitative research. From the basic pricing model (stocks, futures, derivatives, etc.), to alternative Alpha factor digging, and to smart portfolio (optimization) management, real-time assessment, real-time risk management, real-time intelligent stress testing (risk exposure calculation), etc.

TransQuant

TDH

Sophon

TDC