timeseriesAI
timeseriesAI is an open-source project and GitHub organization focused on making state-of-the-art deep learning for time series and sequential data accessible. Its primary project, tsai, provides models, data utilities, transforms, tutorials, and workflows for classification, regression, forecasting, and imputation.
At a Glance
- Python data scientists and machine-learning engineers
- Academic and applied time-series researchers
- Developers building forecasting and sequence-modeling systems
- Organizations working with sensor, business, scientific, or other sequential data
AI Tools by timeseriesAI
(1)
tsai
Time Series Deep Learning Library
Discussions
No discussions yet
Be the first to start a discussion about timeseriesAI
Latest News
tsai 1.0.1 released with packaging fixes for missing subpackages, an import fix, and CI coverage for wheel contents.
tsai 1.0.0 major release introduced breaking dependency/configuration changes, Python 3.10 minimum, newer PyTorch and fastai requirements, and PyPI-only distribution for new versions.
tsai 0.4.1 released.
tsai 0.4.0 released.
Products & Services
Open-source deep-learning library for time series and sequences, built on PyTorch and fastai. It supports classification, regression, forecasting, imputation, and related time-series workflows, and is installable from PyPI with pip.
Market Position
A focused, open-source PyTorch/fastai library emphasizing deep-learning architectures and practical workflows for time-series and sequence data. Its positioning overlaps with broader time-series toolkits such as sktime and Darts and deep-learning forecasting libraries such as PyTorch Forecasting and GluonTS, while tsai differentiates around a large catalog of neural architectures, fastai-style APIs, notebooks, and support for classification as well as forecasting.
Leadership
Founders
Ignacio Oguiza
Creator and cited author of tsai; GitHub profile identifies him with timeseriesAI and places him in Valencia, Spain. LinkedIn search results describe him as founder and CEO of tsaiLab AI Solutions and creator of the open-source tsai library.
Executive Team
Ignacio Oguiza
Creator and maintainer of timeseriesAI/tsai
Open-source developer and data-science practitioner; author credited in the tsai repository and PyPI metadata. His public GitHub profile associates him with timeseriesAI and lists Valencia, Spain.
Founding Story
The project emerged as a practical deep-learning effort around time-series and sequential-data problems, with the initial vision of making modern research methods easier to use through fastai/PyTorch-based open-source tooling, examples, and tutorials. The earliest PyPI release of tsai was published on April 10, 2020.
Business Model
Revenue Model
tsai is distributed as open-source software through GitHub and PyPI under the Apache-2.0 license. The official materials document pip installation and optional extras rather than paid subscriptions, API usage charges, or commercial license tiers.
Target Markets
- Python data scientists and machine-learning engineers
- Academic and applied time-series researchers
- Developers building forecasting and sequence-modeling systems
- Organizations working with sensor, business, scientific, or other sequential data
- Time-series classification
- Multivariate and univariate regression
- Short- and long-term forecasting
- Time-series imputation
- Research and benchmarking of deep-learning architectures on time-series datasets
- Production prototyping by Python/PyTorch/fastai data-science teams