Embedded Markets, Inc., a financial technology firm, today announced its official public launch following ten months of research and development. The company builds AI agent-based simulation and risk management tools that let professional traders stress-test trading strategies across thousands of alternate market histories, rather than the single realized path that history actually took.
Announced in Winnetka, Illinois, the launch introduces a new approach to trading risk management that applies computational sociology and AI agents to address limitations in traditional backtesting as autonomous AI trading agents create new feedback loops in markets.
Traditional backtesting methods fall short in today's markets. Walk-forward tests replay one realized historical path, while Monte Carlo approaches resample returns but treat market participants as passive noise rather than adaptive agents who learn, react, imitate, and adjust leverage endogenously. Meanwhile, autonomous AI trading agents are introducing new forms of feedback loops, herding, model monoculture, and liquidity shocks that traditional backtests do not capture.
"If you've ever watched a trading strategy get destroyed by a four-sigma event or sold just before a large move up, you know what's missing from the trading software marketplace," said Jonathan Haynes, Ph.D., Founder & CEO of Embedded Markets. "This is why our mission is to create the premier risk management platform for trading any type of asset."
Haynes holds a Ph.D. in Sociology from Stanford University and completed postdoctoral research at the Northwestern Institute on Complex Systems and the Kellogg School of Management. He led data science teams for over a decade and later led a global forecasting, pricing, and analytics team at a pharmaceutical firm before founding Embedded Markets.
Initial products include LeveeBacktest Engine and emPortfolioAnalyzer, both targeted for beta release in Q2 2027.
LeveeBacktest Engine, a backtester powered by AI agents, runs strategies through a multiverse of heterogeneous agent-driven histories with network contagion and threshold cascades. Unlike agent-based models governed by fixed rules, or off-the-shelf LLMs handed a trading prompt, this approach uses LoRA fine-tuning at the model level, so collective behaviors emerge from training rather than role-play.
emPortfolioAnalyzer enables regime-specific correlation analyses and hypothetical position sizing aligned with return targets, drawdown tolerance, and risk budgets. Select components will be open-sourced; sign up for beta notifications at www.embeddedmarkets.com.
The platform applies complexity science, sociology, network theory, and AI-driven market simulations to stress-test trading strategies under alternate histories and counterfactual regimes because markets are social, and models should be too.
About Embedded Markets
Embedded Markets, Inc. builds AI agent-based simulation and risk management tools for professional traders. The company applies complexity science, sociology, network theory, and AI-driven market simulations to stress-test trading strategies under alternate histories and counterfactual regimes because markets are social, and models should be too.
