diff --git a/python/cudf/cudf/_lib/CMakeLists.txt b/python/cudf/cudf/_lib/CMakeLists.txt
index 5d4b5421f16..41a7db2285a 100644
--- a/python/cudf/cudf/_lib/CMakeLists.txt
+++ b/python/cudf/cudf/_lib/CMakeLists.txt
@@ -14,7 +14,6 @@
 
 set(cython_sources
     aggregation.pyx
-    avro.pyx
     binaryop.pyx
     column.pyx
     concat.pyx
diff --git a/python/cudf/cudf/_lib/__init__.py b/python/cudf/cudf/_lib/__init__.py
index 918edb6d3f1..57df6899a22 100644
--- a/python/cudf/cudf/_lib/__init__.py
+++ b/python/cudf/cudf/_lib/__init__.py
@@ -2,7 +2,6 @@
 import numpy as np
 
 from . import (
-    avro,
     binaryop,
     concat,
     copying,
diff --git a/python/cudf/cudf/_lib/avro.pyx b/python/cudf/cudf/_lib/avro.pyx
deleted file mode 100644
index b1759635a36..00000000000
--- a/python/cudf/cudf/_lib/avro.pyx
+++ /dev/null
@@ -1,33 +0,0 @@
-# Copyright (c) 2020-2024, NVIDIA CORPORATION.
-
-from cudf._lib.utils cimport data_from_pylibcudf_io
-
-import pylibcudf as plc
-from pylibcudf.io.types import SourceInfo
-
-
-cpdef read_avro(datasource, columns=None, skip_rows=0, num_rows=-1):
-    """
-    Cython function to call libcudf read_avro, see `read_avro`.
-
-    See Also
-    --------
-    cudf.io.avro.read_avro
-    """
-
-    num_rows = -1 if num_rows is None else num_rows
-    skip_rows = 0 if skip_rows is None else skip_rows
-
-    if not isinstance(num_rows, int) or num_rows < -1:
-        raise TypeError("num_rows must be an int >= -1")
-    if not isinstance(skip_rows, int) or skip_rows < 0:
-        raise TypeError("skip_rows must be an int >= 0")
-
-    return data_from_pylibcudf_io(
-        plc.io.avro.read_avro(
-            SourceInfo([datasource]),
-            columns,
-            skip_rows,
-            num_rows
-        )
-    )
diff --git a/python/cudf/cudf/_lib/utils.pxd b/python/cudf/cudf/_lib/utils.pxd
index 7254db5c43d..623c5064a1a 100644
--- a/python/cudf/cudf/_lib/utils.pxd
+++ b/python/cudf/cudf/_lib/utils.pxd
@@ -11,7 +11,7 @@ from pylibcudf.libcudf.table.table cimport table, table_view
 cdef data_from_unique_ptr(
     unique_ptr[table] c_tbl, column_names, index_names=*)
 cdef data_from_pylibcudf_table(tbl, column_names, index_names=*)
-cdef data_from_pylibcudf_io(tbl_with_meta, column_names = *, index_names = *)
+cpdef data_from_pylibcudf_io(tbl_with_meta, column_names = *, index_names = *)
 cdef data_from_table_view(
     table_view tv, object owner, object column_names, object index_names=*)
 cdef table_view table_view_from_columns(columns) except *
diff --git a/python/cudf/cudf/_lib/utils.pyx b/python/cudf/cudf/_lib/utils.pyx
index 9e5b99f64eb..292de82e4c4 100644
--- a/python/cudf/cudf/_lib/utils.pyx
+++ b/python/cudf/cudf/_lib/utils.pyx
@@ -316,7 +316,7 @@ cdef data_from_pylibcudf_table(tbl, column_names, index_names=None):
         index_names
     )
 
-cdef data_from_pylibcudf_io(tbl_with_meta, column_names=None, index_names=None):
+cpdef data_from_pylibcudf_io(tbl_with_meta, column_names=None, index_names=None):
     """
     Unpacks the TableWithMetadata from libcudf I/O
     into a dict of columns and an Index (cuDF format)
diff --git a/python/cudf/cudf/io/avro.py b/python/cudf/cudf/io/avro.py
index 964bd02b03e..11730e98c95 100644
--- a/python/cudf/cudf/io/avro.py
+++ b/python/cudf/cudf/io/avro.py
@@ -1,7 +1,9 @@
 # Copyright (c) 2019-2024, NVIDIA CORPORATION.
 
+import pylibcudf as plc
+
 import cudf
-from cudf import _lib as libcudf
+from cudf._lib.utils import data_from_pylibcudf_io
 from cudf.utils import ioutils
 
 
@@ -23,8 +25,19 @@ def read_avro(
         filepath_or_buffer, "read_avro"
     )
 
-    return cudf.DataFrame._from_data(
-        *libcudf.avro.read_avro(
-            filepath_or_buffer, columns, skiprows, num_rows
-        )
+    num_rows = -1 if num_rows is None else num_rows
+    skip_rows = 0 if skiprows is None else skiprows
+
+    if not isinstance(num_rows, int) or num_rows < -1:
+        raise TypeError("num_rows must be an int >= -1")
+    if not isinstance(skip_rows, int) or skip_rows < 0:
+        raise TypeError("skip_rows must be an int >= 0")
+
+    plc_result = plc.io.avro.read_avro(
+        plc.io.types.SourceInfo([filepath_or_buffer]),
+        columns,
+        skip_rows,
+        num_rows,
     )
+
+    return cudf.DataFrame._from_data(*data_from_pylibcudf_io(plc_result))