Strategy research

MFT can either help a client to build tailor-made investment strategies using machine learning or improve existing one. In our research we use assembly-line, meta-strategy appoach where each individual is responsible for a separate part of strategy development: data prepocessing, labelling, feature generation and analysis, cross-validation and backtesting.


MFT provides various consulting services regarding machine learning in investment strategies development. Our services inlcude assistance in labeling, feature generation, cross-validation, model training, position sizing and backtesting.

Strategy backtest overfit detection

We help investment companies to build the framework detecting false discoveries. This makes strategy development more robust and sustainable out-of-sample.

Machine Learning Asset Allocation

MFT helps investment managers construct robust investment portfolios using various supervised and unsupervised machine learning techniques. Portfolio algorithms can be customized based on client's fitness function: Sharpe ratio, Sortino ratio, Maximum Daily Drawdown, etc.