diff --git a/assignment.md b/assignment.md
new file mode 100644
index 0000000..ed0427d
--- /dev/null
+++ b/assignment.md
@@ -0,0 +1,14 @@
+# Technical assignment
+
+The multi tenant application found in this repository is a striped down but functional example of what we need from our production system.
+The current state of this application has some shortcomings, that stops us from choosing this data storage architecture:
+
+- with each organization living in its own database in a specific database server, we need to replicate all user data into each organization specific database.
+  We do not need to sync all users into all organization's dbs, just the users, who are part of a given organization
+- each organization has to publish their resource data into the main database and the main database has to subscribe to them.
+  In the long run the number of active subscriptions in the main database scales 1:1 with the number of organizations, which is not ideal.
+- Each organization may have access to multiple regions. Each organization's region currently means 1 `knex` database connection client with a connection pool.
+  In a production workload, the connection pool would use 10+ connections. We'd run multiple instances (~ 10 pods) of the back-end server.
+  Given a 100 organizations, having access to 3 regions each, the number of active connections that are targeting each database server becomes huge. This is not scalable.
+- Each organizations database connection's are now kept in an in memory map, that can result storing a lot not used clients in memory.
+  How can we optimize for efficient but relatively fast client access?