“We shape our tools, and thereafter our tools shape us.”

― Marshall McLuhan


Artifact Research maintains both open-source and commercial software products that facilitate the analysis of small-data problems and time series data, as well as tools for financial risk mitigation. We focus on building lean Python frameworks that can be easily customized and extended. If you wish to integrate our products into your workflow, do not hesitate to contact us!

Open-source frameworks

The following framework is licensed under the permissive open-source MIT license, free to use for personal and commercial projects.

bayesloop is a Probabilistic programming framework that facilitates objective model selection for time-varying parameter models. It serves a specific niche of statistical models: We focus on well-known, simple statistical models with only few parameters, and extend these parameters to be time-varying. Its main advantage over other inference frameworks is that it can apply Occam’s razor to decide how model parameters change over time.

Artifact Research supports this open-source Python project as it is actively used in many of our projects. bayesloop has also been applied in scientific studies (see here and here) on metastatic cancer cells, anomaly detection in financial time series, the comparison of climate models, and accident analysis.

Analyze time series data either retrospectively or in real-time to see how parameters/indicators change

Test hypotheses about how model parameters change, e.g. whether parameters drift slowly or change abruptly in regime switches or structural breaks

Automatically handle missing data points and fill in the parameter evolution at these time steps

Full online documentation available, including step-by-step tutorials and example applications

Commercial frameworks

For details on licensing options and pricing of our commercial software products, please contact us.

TYCHO is a Python framework for long-term portfolio optimization. Rather than reducing candidate portfolios to a set of expected return values and a covariance matrix, it directly simulates the performance of frequently re-balanced candidate portfolios, thus accounting for non-linear correlations between assets, auto-correlated returns, and higher moments of the return distribution.

This simulation approach is combined with robust optimization heuristics, allowing the user to optimize exactly for the objectives they care for the most. Minimize for example the magnitude or duration of draw-downs, maximize the positive skewness of portfolio returns, maximize ending wealth, or add your own objective with just a few lines of Python code.

Accurately simulate the interplay of financial assets over time and the effect on long-term compounded returns

Optimize re-balancing or re-allocation periods based on transaction costs

Test the effect of periodic withdrawals on portfolio performance and draw-downs

Optimize for the objectives that really matter to you or your clients, such as draw-downs and ending wealth

Include highly non-linear products such as a rolling options strategy to portfolios to discover efficient hedges and optimize their allocation

TYCHO is currently in BETA stage. Subscribe to our newsletter to receive updates on future releases!

TENTACLE is a Python framework to discover and deploy hedged investment strategies that trade cryptocurrencies and derivatives on the Kraken exchange and its Futures exchange.

It provides a range of building blocks that accelerate the iteration process from strategy development to production, including ready-to-use code for data collection, strategy simulation, efficient order execution, transfers between spot and derivative markets, and monitoring of live trading strategies.

While we certainly do not advocate cryptocurrencies as the future of finance, and we acknowledge that any active investment in cryptocurrencies carries significant counterparty risk, we also believe that these markets currently offer attractive inefficiencies that can be exploited by hedged strategies without being exposed to the high volatility inherent to the underlying assets.

TENTACLE is currently the basis for multiple trading strategies that focus on two main mechanisms:

First, TENTACLE allows to seamlessly execute basis trades that involve both spot currencies and derivatives such as perpetual futures contracts. By trading on both the spot- and the futures-exchange simultaneously, TENTACLE minimizes holding periods of unhedged positions, and combines basis trades with additional strategies such as momentum strategies.

Second, TENTACLE specifically targets Kraken’s backstop liquidity provider program for Futures contracts. This allows investors to receive liquidated positions at very favorable prices. TENTACLE provides end-to-end automation for this process.

If you are interested in using TENTACLE for your trading system or you are interested in one of the strategies detailed above, please contact us!