Simudyne

Simudyne develops enterprise simulation software built around agent-based modeling, allowing organizations to construct virtual environments in which large numbers of individual entities interact so that complex system dynamics — contagion effects, feedback loops and structural change — can be reproduced and studied. The platform is positioned for problems that the company says traditional modeling methods handle poorly, including risk management, market execution and financial crime analytics. Customers described by the company include global banks, regulators and exchanges.

The current product set is organized in three layers: Simudyne Horizon, an agentic process for automated model discovery, calibration and validation intended to shorten the path from research to production; Simudyne Pulse, foundation and multi-agent simulation models that reproduce real-world system dynamics; and Simudyne Nexus, environments used to train reasoning models to explore and optimize systems. Models are validated through a six-step calibration process meant to ensure that design and parameters reproduce statistical and behavioral dynamics matching real-world data. Results are surfaced through customizable, interactive dashboards and a web console generated from code annotations, and output data can be exported in formats such as Parquet for downstream data-science workflows.

An earlier generation of the platform was marketed as Providence, described in 2017 as software that let banks build AI-driven simulations of complex systems such as the US bond market and the UK housing market, the latter tracking roughly 27 million households to assess the effects of policy levers such as loan-to-value and loan-to-income ratios.

Founding story

Simudyne was founded in 2016 by Justin Lyon, a simulation specialist who began working for the US Department of Defense in Afghanistan and Iraq in 2001 simulating insurgencies and counter-insurgencies, then built simulations for gas, oil and insurance companies and spent five years working for the Bank of England. He has said the motivation for the business came from dissatisfaction with how leaders make decisions, particularly the reliance on averages exposed by the 2008 financial crisis.

Business model

Simudyne licenses simulation software to enterprises, primarily in financial services. The software is described as a deployed solution that sits behind the customer's firewall so that client data and models remain proprietary, and it includes a library of reusable code for frequently used and specialized simulation functions.

Traction

The company reported a 600% year-on-year increase in revenue in 2018 following the addition of new global banking clients, and a headcount of 30 staff members including six PhDs with expertise spanning market simulation, fraud detection and risk management. It has stated that its client base includes global banks, regulators and exchanges, and that it signed a commercial deal with Barclays.

Latest developments

The company's platform is currently presented in three components — Simudyne Horizon for automated model discovery, calibration and validation; Simudyne Pulse foundation and multi-agent simulation models; and Simudyne Nexus environments for training reasoning models — alongside customizable, interactive dashboards for real-time analysis.