I have multiple tests in one file. I want to create users at the beginning of each test and roll back those db changes at the end of each test.
At the top of my file I have this:

  use Teiserver.DataCase, async: true

which references this module

defmodule Teiserver.DataCase do


  use ExUnit.CaseTemplate

  using do
    quote do
      alias Teiserver.Repo

      import Ecto
      import Ecto.Changeset
      import Ecto.Query
      import Teiserver.DataCase
    end
  end

  @doc """
  Sets up the sandbox based on the test tags.
  """
  def setup_sandbox(tags) do
    pid = Ecto.Adapters.SQL.Sandbox.start_owner!(Teiserver.Repo, shared: not tags[:async])
    on_exit(fn -> Ecto.Adapters.SQL.Sandbox.stop_owner(pid) end)
  end

  setup tags do
    setup_sandbox(tags)
    :ok
  end

  @doc """
  A helper that transforms changeset errors into a map of messages.

      assert {:error, changeset} = Accounts.create_user(%{password: "short"})
      assert "password is too short" in errors_on(changeset).password
      assert %{password: ["password is too short"]} = errors_on(changeset)

  """
  def errors_on(changeset) do
    Ecto.Changeset.traverse_errors(changeset, fn {message, opts} ->
      Regex.replace(~r"%{(\w+)}", message, fn _, key ->
        opts |> Keyword.get(String.to_existing_atom(key), key) |> to_string()
      end)
    end)
  end
end

However it seems like my changes are persisted between tests.

Also test_helper.exs has this

ExUnit.start()
Ecto.Adapters.SQL.Sandbox.mode(Teiserver.Repo, :manual)

alias Teiserver.Repo
alias Central.Helpers.GeneralTestLib

:ok = Ecto.Adapters.SQL.Sandbox.checkout(Repo, sandbox: false)

if not GeneralTestLib.seeded?() do
  GeneralTestLib.seed()
  Teiserver.TeiserverTestLib.seed()
end

Ecto.Adapters.SQL.Sandbox.checkin(Repo)

How do I make it so any db changes are rolled back after each test?

You can see the video in this link where I run the same test multiple times and it will just randomly fail eventually.