diff --git a/README.rst b/README.rst
index 3511d1ca91..ce7fd3c7a0 100644
--- a/README.rst
+++ b/README.rst
@@ -14,7 +14,11 @@ If you'll be modifying ORA code, the `Open edX Developer Stack`_ is a Docker-bas
 edX engineers follow the `guides on our wiki <https://openedx.atlassian.net/wiki/spaces/EDUCATOR/pages/9765004/ORA+Developer+Guide>`_.
 Reading this page before contributing **is highly recommended**.
 
+Additionally, consult the `Repo Docs`_ and `ORA Developer Guide`_ for architectural decisions and other information on how to develop for ORA.
+
 .. _Open edX Developer Stack: https://github.com/openedx/devstack
+.. _Repo Docs: https://github.com/openedx/edx-ora2/tree/master/docs
+.. _ORA Developer Guide: https://github.com/openedx/edx-ora2/tree/master/docs/developers_guide.rst
 
 License
 =======
diff --git a/docs/decisions/0003-ui-mixins.rst b/docs/decisions/0003-ui-mixins.rst
new file mode 100644
index 0000000000..db93f6b30c
--- /dev/null
+++ b/docs/decisions/0003-ui-mixins.rst
@@ -0,0 +1,65 @@
+#######################
+0003 UI Mixins
+#######################
+
+Status
+******
+
+Accepted (September 2023)
+
+Context
+*******
+
+Previously, ORA frontend development has been challenging. ORA frontends were built
+using a combination of a placeholder Django template and injection of per-step HTML.
+Additionally, the way we built the context and HTML per section mixed business and
+presentation logic in complicated ways. This made refactoring and accessing that
+underlying data in any other context challenging.
+
+Along the way to creating a new ORA UI, we decided on some refactoring to aid access
+to that data and allow us to reuse that code across UI implementations.
+
+Decisions
+*********
+
+- We will create standalone APIs for accessing data on an ORA block, placed in
+  ``openassessment/xblock/apis``. These APIs will be instantiated at block load and act
+  as the primary way we access data / functions on the ORA block.
+- We will move existing presentation code into ``openassessment/xblock/ui_mixins/legacy``,
+  refactoring to leverage our newly-created APIs for data access. 
+- New UI variants will also be created in ``openassessment/xblock/ui_mixins`` as separate
+  folders.
+
+Consequences
+************
+
+Separate data / utility APIs
+==============================
+
+Splitting the data / utilities out of existing mixins and into a separate location
+allows us to access that data across multiple contexts, without having to duplicate
+business logic.
+
+These APIs, in addition to serving our legacy UI views, can also be leveraged in future
+UI views or other contexts that need access to that data or functions.
+
+Having those APIs and business logic in a single place will help us keep behavior from
+unintentionally drifting between UIs or other endpoints.
+
+Refactoring legacy presentation code
+====================================
+
+Refactoring our legacy presentation will allow us to leverage and exercise our APIs
+ensuring behavior is not altered from before refactor to after refactor.
+
+Moving this legacy presentation code will also allow us to decouple it from its
+currently tightly-coupled location inside of the openassessment block.
+
+Location for new UI variants
+============================
+
+Creating a specific location for UI variants (past and present) helps clearly
+separate our presentation and business concerns.
+
+Additionally, this gives us a well-defined pattern for introducing new UIs or iterating
+on our UIs moving forward.