Tide Framework
Built for the scale of high-frequency data and the complexity of financial markets, Tide is Lake Shore Dynamics' signal research framework for deep learning on market data. Tide handles the engineering beneath the model layer so researchers can focus on signal discovery, and supports order-level market simulation during training.
Data Infrastructure
Synchronize and stream data occurring at different timescales into a coherent training pipeline, without full-dataset preprocessing. Tide coordinates data occurring at different frequencies, enforcing temporal and causal structure as it is fed to models.
Experiment Management
Configuration-driven experimentation. Every model-dataset pairing and run is tracked by a built-in registry. Full reproducibility out of the box. Checkpointing, logging and export to dashboards for visibility on experiments.
Training and Evaluation
High-frequency market simulation integrated directly into the training loop. Model and strategy performance is easily monitored throughout training, surfacing degradation, regime sensitivity and overfitting as it happens.
Deployment
Lake Shore Dynamics adapts Tide to your dataset(s) and offers continuous support and development in the application of state-of-the-art models to your specific objective.
Applications
Forecasting
Reinforcement learning
Regime detection
Optimal execution
Philosophy
We founded Lake Shore Dynamics on our conviction that it is both necessary and possible to model financial markets from the level of individual orders.
The behavior of market participants is the ultimate filter for the information involved in price and market formation. If we can explain markets at the resolution of order flow, not only can we realistically assess strategies: we can model the impact of macroeconomic events and company-specific news that affect markets and our ability to trade successfully.
Models invented in this decade alone have made massive strides in human reasoning and language, image generation and decoding the basis of life. Building on this paradigm, we are developing our own architectures to model market behavior.
We are researchers and builders with backgrounds in mathematics, statistics and computer science. We default to skepticism, whether evaluating our own results or those of others. We believe that large-scale market data and architectural innovation will allow us to capture the idiosyncratic behavior of market participants, without imposing biasing assumptions.
By training models on both order flow and the broader market, our framework will allow us to interrogate cause and effect in markets on every dimension. We are modeling the fabric of markets and the conditions under which it is woven.
Team
