diff --git a/doc/development.rst b/doc/development.rst
new file mode 100644
index 0000000..a4a1cc9
--- /dev/null
+++ b/doc/development.rst
@@ -0,0 +1,74 @@
+.. _development:
+
+PEtab SciML development process
+===============================
+
+This page describes the development process for PEtab SciML, an extension to the
+`PEtab format <https://petab.readthedocs.io/>`_ for parameter estimation of dynamic models.
+
+PEtab SciML format development process
+--------------------------------------
+
+We know PEtab SciML cannot accommodate every use case, but we are committed to addressing
+current  and future requirements in upcoming versions of the format. We value backwards
+compatibility, but do not exclude breaking changes when justified by clear benefits.
+
+Anyone is welcome to propose changes or additions to PEtab SciML. Proposals should be made
+via GitHub issues, where benefits, potential problems, and alternatives can be discussed.
+
+A proposal is considered accepted for inclusion in the next PEtab SciML version if it is
+endorsed by a majority of the maintainers and at least one tool supporting PEtab SciML
+provides a prototype implementation. For accepted changes, corresponding test cases covering
+the change must be added to the
+`PEtab SciML test suite <https://github.com/sebapersson/petab_sciml_testsuite>`_
+before release.
+
+Requirements for new releases:
+
+- Updated format specification
+- Updated validator (linter)
+- Updated test suite
+- Updated changelog
+
+For each release, the core maintainers ensure that
+
+- a new release is created in the GitHub repository
+- the new version of the specification is deposited on Zenodo
+
+Versioning of PEtab SciML
+-------------------------
+
+The PEtab SciML specification follows `semantic versioning <https://semver.org/>`_. Changes
+to the specification require a new release. Necessary clarifications or corrections may be
+collected on an Errata page until the next release.
+
+Release timing is at the discretion of the maintainers. However, accepted changes should be
+released within two months of acceptance.
+
+Maintainers
+-----------
+
+PEtab SciML, like PEtab, is intended as a community effort. Decisions should, as far as
+possible, be made with input from the broader community. Nevertheless, PEtab SciML has a
+set of core maintainers responsible for maintaining the software and specification, and
+for approving changes.
+
+New maintainers are welcome, but a track record is expected (e.g., contributions to PEtab
+SciML or other standards such as PEtab, SBML, etc.).
+
+Values
+------
+
+We are committed to diversity, open communication, transparent processes, and a welcoming
+environment. We aim for clear processes without unnecessary over-formalization that could
+slow development.
+
+Communication channels
+----------------------
+
+The primary communication channels are the GitHub
+`issues <https://github.com/PEtab-dev/petab_sciml/issues>`_ and
+`discussions <https://github.com/PEtab-dev/petab_sciml/discussions>`_ pages.
+Additionally, since PEtab SciML is a PEtab extension, the PEtab
+`mailing list <https://groups.google.com/g/petab-discuss>`_ can be used for discussions
+that are not well suited for GitHub.
diff --git a/doc/index.rst b/doc/index.rst
index 35d0d90..519ab34 100644
--- a/doc/index.rst
+++ b/doc/index.rst
@@ -26,7 +26,7 @@ PEtab SciML - Scientific Machine Learning Format and Tooling
 
    format
    layers
-   tutorial
+   development
 
 .. toctree::
    :caption: Python package
diff --git a/doc/tutorial.rst b/doc/tutorial.rst
deleted file mode 100644
index ac51ab0..0000000
--- a/doc/tutorial.rst
+++ /dev/null
@@ -1,14 +0,0 @@
-Tutorial
-========
-
-The tutorials will consist of:
-
-1. Overarching tutorial where we show how to implement the
-   Lotka-Volterra problem from the UDE paper.
-2. Extended tutorial 1 where we have a neural-network setting
-   parameters, here it is also worthwhile to consider simulation
-   conditions.
-3. Extended tutorial 2 where we have a neural network in the observable
-   function.
-
-If time, add tutorial for having two neural networks.