diff --git a/cpp/include/cuvs/cluster/agglomerative.hpp b/cpp/include/cuvs/cluster/agglomerative.hpp
index 817826856a..d80086e218 100644
--- a/cpp/include/cuvs/cluster/agglomerative.hpp
+++ b/cpp/include/cuvs/cluster/agglomerative.hpp
@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2021-2025, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 
@@ -129,7 +129,7 @@ struct distance_params {
 /** Specialized parameters to build the Mutual Reachability graph */
 struct mutual_reachability_params {
   /** this neighborhood will be selected for core distances. */
-  int min_samples;
+  int min_samples = 5;
 
   /** weight applied when internal distance is chosen for mutual reachability (value of 1.0 disables
    * the weighting) */
diff --git a/cpp/include/cuvs/distance/distance.hpp b/cpp/include/cuvs/distance/distance.hpp
index 13c8c7bd7e..b5454f00a3 100644
--- a/cpp/include/cuvs/distance/distance.hpp
+++ b/cpp/include/cuvs/distance/distance.hpp
@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2021-2025, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 
@@ -93,10 +93,10 @@ enum KernelType { LINEAR, POLYNOMIAL, RBF, TANH };
  */
 struct KernelParams {
   // Kernel function parameters
-  KernelType kernel;  //!< Type of the kernel function
-  int degree;         //!< Degree of polynomial kernel (ignored by others)
-  double gamma;       //!< multiplier in the
-  double coef0;       //!< additive constant in poly and tanh kernels
+  KernelType kernel = KernelType::LINEAR;  //!< Type of the kernel function
+  int degree        = 3;                   //!< Degree of polynomial kernel (ignored by others)
+  double gamma      = 1.0;                 //!< multiplier in the
+  double coef0      = 0.0;                 //!< additive constant in poly and tanh kernels
 };
 }  // end namespace kernels
 
diff --git a/cpp/include/cuvs/distance/grammian.hpp b/cpp/include/cuvs/distance/grammian.hpp
index 4cb3aa47b6..a179ae80f8 100644
--- a/cpp/include/cuvs/distance/grammian.hpp
+++ b/cpp/include/cuvs/distance/grammian.hpp
@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2022-2024, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2022-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 
@@ -36,15 +36,17 @@ using csr_input_matrix_view_t = raft::device_csr_matrix_view<const math_t, int,
 template <typename math_t>
 class GramMatrixBase {
  protected:
-  cublasHandle_t cublas_handle;
+  cublasHandle_t cublas_handle = nullptr;
   bool legacy_interface;
 
  public:
-  GramMatrixBase() : legacy_interface(false) {};
+  GramMatrixBase() : legacy_interface(false) {}
   [[deprecated]] GramMatrixBase(cublasHandle_t cublas_handle)
-    : cublas_handle(cublas_handle), legacy_interface(true) {};
+    : cublas_handle(cublas_handle), legacy_interface(true)
+  {
+  }
 
-  virtual ~GramMatrixBase() {};
+  virtual ~GramMatrixBase() = default;
 
   /** Convenience function to evaluate the Gram matrix for two vector sets.
    *  Vector sets are provided in Matrix format
@@ -320,10 +322,14 @@ class PolynomialKernel : public GramMatrixBase<math_t> {
    * @param offset
    */
   PolynomialKernel(exp_t exponent, math_t gain, math_t offset)
-    : GramMatrixBase<math_t>(), exponent(exponent), gain(gain), offset(offset) {};
+    : GramMatrixBase<math_t>(), exponent(exponent), gain(gain), offset(offset)
+  {
+  }
 
   [[deprecated]] PolynomialKernel(exp_t exponent, math_t gain, math_t offset, cublasHandle_t handle)
-    : GramMatrixBase<math_t>(handle), exponent(exponent), gain(gain), offset(offset) {};
+    : GramMatrixBase<math_t>(handle), exponent(exponent), gain(gain), offset(offset)
+  {
+  }
 
   /** Evaluate kernel matrix using polynomial kernel.
    *
@@ -436,7 +442,9 @@ class TanhKernel : public GramMatrixBase<math_t> {
   TanhKernel(math_t gain, math_t offset) : GramMatrixBase<math_t>(), gain(gain), offset(offset) {}
 
   [[deprecated]] TanhKernel(math_t gain, math_t offset, cublasHandle_t handle)
-    : GramMatrixBase<math_t>(handle), gain(gain), offset(offset) {};
+    : GramMatrixBase<math_t>(handle), gain(gain), offset(offset)
+  {
+  }
 
   /** Evaluate kernel matrix using tanh kernel.
    *
@@ -551,10 +559,12 @@ class RBFKernel : public GramMatrixBase<math_t> {
    * @tparam math_t floating point type
    * @param gain
    */
-  RBFKernel(math_t gain) : GramMatrixBase<math_t>(), gain(gain) {};
+  RBFKernel(math_t gain) : GramMatrixBase<math_t>(), gain(gain) {}
 
   [[deprecated]] RBFKernel(math_t gain, cublasHandle_t handle)
-    : GramMatrixBase<math_t>(handle), gain(gain) {};
+    : GramMatrixBase<math_t>(handle), gain(gain)
+  {
+  }
 
   void matrixRowNormL2(raft::resources const& handle,
                        dense_input_matrix_view_t<math_t> matrix,
diff --git a/cpp/include/cuvs/neighbors/ball_cover.hpp b/cpp/include/cuvs/neighbors/ball_cover.hpp
index 17fa94bfd7..039df42672 100644
--- a/cpp/include/cuvs/neighbors/ball_cover.hpp
+++ b/cpp/include/cuvs/neighbors/ball_cover.hpp
@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 
@@ -135,7 +135,7 @@ struct index : cuvs::neighbors::index {
   raft::device_matrix<float, int64_t, raft::row_major> X_reordered;
 
  protected:
-  bool index_trained;
+  bool index_trained = false;
 };
 
 /** @} */
diff --git a/cpp/include/cuvs/neighbors/common.hpp b/cpp/include/cuvs/neighbors/common.hpp
index 39967999ed..9d99b800fb 100644
--- a/cpp/include/cuvs/neighbors/common.hpp
+++ b/cpp/include/cuvs/neighbors/common.hpp
@@ -502,7 +502,7 @@ namespace filtering {
 enum class FilterType { None, Bitmap, Bitset };
 
 struct base_filter {
-  ~base_filter()                             = default;
+  virtual ~base_filter()                     = default;
   virtual FilterType get_filter_type() const = 0;
 };
 
@@ -976,7 +976,7 @@ struct mg_index {
   auto operator=(mg_index&&) -> mg_index&      = default;
 
   distribution_mode mode_;
-  int num_ranks_;
+  int num_ranks_ = 0;
   std::vector<iface<AnnIndexType, T, IdxT>> ann_interfaces_;
 
   // for load balancing mechanism
diff --git a/cpp/include/cuvs/neighbors/hnsw.hpp b/cpp/include/cuvs/neighbors/hnsw.hpp
index a1be1c58cc..f58f772294 100644
--- a/cpp/include/cuvs/neighbors/hnsw.hpp
+++ b/cpp/include/cuvs/neighbors/hnsw.hpp
@@ -738,9 +738,9 @@ void extend(raft::resources const& res,
  */
 
 struct search_params : cuvs::neighbors::search_params {
-  int ef;               // size of the candidate list
-  int num_threads = 0;  // number of host threads to use for concurrent searches. Value of 0
-                        // automatically maximizes parallelism
+  int ef          = 200;  // size of the candidate list
+  int num_threads = 0;    // number of host threads to use for concurrent searches. Value of 0
+                          // automatically maximizes parallelism
 };
 
 /**
diff --git a/cpp/include/cuvs/neighbors/vamana.hpp b/cpp/include/cuvs/neighbors/vamana.hpp
index 6c0fc94c1e..224d3bd97d 100644
--- a/cpp/include/cuvs/neighbors/vamana.hpp
+++ b/cpp/include/cuvs/neighbors/vamana.hpp
@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2024-2025, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 
@@ -33,8 +33,8 @@ namespace cuvs::neighbors::vamana {
  */
 template <typename T = float>
 struct codebook_params {
-  int pq_codebook_size;
-  int pq_dim;
+  int pq_codebook_size = 0;
+  int pq_dim           = 0;
   std::vector<T> pq_encoding_table;
   std::vector<T> rotation_matrix;
 };
diff --git a/cpp/include/cuvs/preprocessing/spectral_embedding.hpp b/cpp/include/cuvs/preprocessing/spectral_embedding.hpp
index e7a578d2ab..5575b1bb38 100644
--- a/cpp/include/cuvs/preprocessing/spectral_embedding.hpp
+++ b/cpp/include/cuvs/preprocessing/spectral_embedding.hpp
@@ -24,10 +24,10 @@ namespace cuvs::preprocessing::spectral_embedding {
  */
 struct params {
   /** @brief The number of components to reduce the data to. */
-  int n_components;
+  int n_components = 2;
 
   /** @brief The number of neighbors to use for the nearest neighbors graph. */
-  int n_neighbors;
+  int n_neighbors = 15;
 
   /**
    * @brief Whether to normalize the Laplacian matrix.
@@ -36,7 +36,7 @@ struct params {
    * If false, uses the unnormalized graph Laplacian (L = D - W).
    * Normalized Laplacian often leads to better results for clustering tasks.
    */
-  bool norm_laplacian;
+  bool norm_laplacian = true;
 
   /**
    * @brief Whether to drop the first eigenvector.
@@ -45,7 +45,7 @@ struct params {
    * uninformative. Setting this to true drops it from the embedding.
    * This is typically set to true when norm_laplacian is true.
    */
-  bool drop_first;
+  bool drop_first = true;
 
   /**
    * @brief Tolerance for the eigenvalue solver.