About Me¶
Marcos Manuel Muraro is the author of Jano.
The project grew out of a practical need: evaluating systems whose behavior changes across time requires more than a random split. Jano is meant to make those simulations explicit, inspectable and easier to communicate.
You can connect with me here:
LinkedIn: https://www.linkedin.com/in/marmurar/
Project direction¶
Jano is currently focused on:
clear temporal partition definitions,
support for pandas, NumPy and Polars inputs,
efficient splitting for larger datasets,
simulation reporting that is useful before model training begins.
Citation¶
If you use Jano in research, technical reports, benchmarks or production validation work, please cite the project with this BibTeX entry:
@software{muraro_jano_2026,
author = {Muraro, Marcos Manuel},
title = {Jano: Temporal Simulation and Backtesting Toolkit for Time-Dependent Machine Learning Systems},
year = {2026},
version = {0.4.1},
url = {https://github.com/marmurar/jano},
repository = {https://github.com/marmurar/jano},
doi = {10.5281/zenodo.20301006},
license = {MIT},
note = {Python toolkit for temporal simulation, walk-forward validation, backtesting, and retraining-policy analysis for time-dependent machine learning systems}
}
The same metadata is available in CITATION.cff. The archived Zenodo record is
available at https://doi.org/10.5281/zenodo.20301006.