diff --git a/sql-plugin/src/main/320+/scala/com/nvidia/spark/rapids/shims/v2/Spark32XShims.scala b/sql-plugin/src/main/320+/scala/com/nvidia/spark/rapids/shims/v2/Spark32XShims.scala
index 27bbc361f19..907d608af74 100644
--- a/sql-plugin/src/main/320+/scala/com/nvidia/spark/rapids/shims/v2/Spark32XShims.scala
+++ b/sql-plugin/src/main/320+/scala/com/nvidia/spark/rapids/shims/v2/Spark32XShims.scala
@@ -642,8 +642,10 @@ trait Spark32XShims extends SparkShims {
             + TypeSig.ARRAY).nested(), TypeSig.all),
         (scan, conf, p, r) => new SparkPlanMeta[InMemoryTableScanExec](scan, conf, p, r) {
           override def tagPlanForGpu(): Unit = {
-            if (!scan.relation.cacheBuilder.serializer.isInstanceOf[ParquetCachedBatchSerializer]) {
-              willNotWorkOnGpu("ParquetCachedBatchSerializer is not being used")
+            scan.relation.cacheBuilder.serializer match {
+              case _: com.nvidia.spark.ParquetCachedBatchSerializer => ()
+              case _ =>
+                willNotWorkOnGpu("ParquetCachedBatchSerializer is not being used")
             }
           }
 
@@ -885,7 +887,7 @@ trait Spark32XShims extends SparkShims {
       exportColumnRdd: Boolean): GpuColumnarToRowExecParent = {
     val serName = plan.conf.getConf(StaticSQLConf.SPARK_CACHE_SERIALIZER)
     val serClass = ShimLoader.loadClass(serName)
-    if (serClass == classOf[ParquetCachedBatchSerializer]) {
+    if (serClass == classOf[com.nvidia.spark.ParquetCachedBatchSerializer]) {
       GpuColumnarToRowTransitionExec(plan)
     } else {
       GpuColumnarToRowExec(plan)