diff --git a/cmd/suggestion/random/v1alpha2/Dockerfile b/cmd/suggestion/random/v1alpha2/Dockerfile
new file mode 100644
index 00000000000..734e7cdff91
--- /dev/null
+++ b/cmd/suggestion/random/v1alpha2/Dockerfile
@@ -0,0 +1,8 @@
+FROM python:3
+
+ADD . /usr/src/app/github.com/kubeflow/katib
+WORKDIR /usr/src/app/github.com/kubeflow/katib/cmd/suggestion/random/v1alpha2
+RUN pip install --no-cache-dir -r requirements.txt
+ENV PYTHONPATH /usr/src/app/github.com/kubeflow/katib:/usr/src/app/github.com/kubeflow/katib/pkg/api/v1alpha2/python
+
+ENTRYPOINT ["python", "main.py"]
diff --git a/cmd/suggestion/random/v1alpha2/__init__.py b/cmd/suggestion/random/v1alpha2/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/cmd/suggestion/random/v1alpha2/main.py b/cmd/suggestion/random/v1alpha2/main.py
new file mode 100644
index 00000000000..26663eac480
--- /dev/null
+++ b/cmd/suggestion/random/v1alpha2/main.py
@@ -0,0 +1,23 @@
+import grpc
+import time
+from pkg.api.v1alpha2.python import api_pb2_grpc
+from pkg.suggestion.v1alpha2.random_service import RandomService
+from concurrent import futures
+
+_ONE_DAY_IN_SECONDS = 60 * 60 * 24
+DEFAULT_PORT = "0.0.0.0:6789"
+
+def serve():
+    server = grpc.server(futures.ThreadPoolExecutor(max_workers=10))
+    api_pb2_grpc.add_SuggestionServicer_to_server(RandomService(), server)
+    server.add_insecure_port(DEFAULT_PORT)
+    print("Listening...")
+    server.start()
+    try:
+        while True:
+            time.sleep(_ONE_DAY_IN_SECONDS)
+    except KeyboardInterrupt:
+        server.stop(0)
+
+if __name__ == "__main__":
+    serve()
diff --git a/cmd/suggestion/random/v1alpha2/requirements.txt b/cmd/suggestion/random/v1alpha2/requirements.txt
new file mode 100644
index 00000000000..8d2c9d4bda7
--- /dev/null
+++ b/cmd/suggestion/random/v1alpha2/requirements.txt
@@ -0,0 +1,9 @@
+grpcio
+duecredit
+cloudpickle==0.5.6
+numpy>=1.13.3
+scikit-learn>=0.19.0
+scipy>=0.19.1
+forestci
+protobuf
+googleapis-common-protos
diff --git a/manifests/v1alpha2/katib/suggestion/random/deployment.yaml b/manifests/v1alpha2/katib/suggestion/random/deployment.yaml
new file mode 100644
index 00000000000..af7093548da
--- /dev/null
+++ b/manifests/v1alpha2/katib/suggestion/random/deployment.yaml
@@ -0,0 +1,23 @@
+apiVersion: extensions/v1beta1
+kind: Deployment
+metadata:
+  name: katib-suggestion-random
+  namespace: kubeflow
+  labels:
+    app: katib
+    component: suggestion-random
+spec:
+  replicas: 1
+  template:
+    metadata:
+      name: katib-suggestion-random
+      labels:
+        app: katib
+        component: suggestion-random
+    spec:
+      containers:
+      - name: katib-suggestion-random
+        image: katib/suggestion-random
+        ports:
+        - name: api
+          containerPort: 6789
diff --git a/manifests/v1alpha2/katib/suggestion/random/service.yaml b/manifests/v1alpha2/katib/suggestion/random/service.yaml
new file mode 100644
index 00000000000..653db93aa1c
--- /dev/null
+++ b/manifests/v1alpha2/katib/suggestion/random/service.yaml
@@ -0,0 +1,17 @@
+apiVersion: v1
+kind: Service
+metadata:
+  name: katib-suggestion-random
+  namespace: kubeflow
+  labels:
+    app: katib
+    component: suggestion-random
+spec:
+  type: ClusterIP
+  ports:
+    - port: 6789
+      protocol: TCP
+      name: api
+  selector:
+    app: katib
+    component: suggestion-random
diff --git a/pkg/suggestion/v1alpha2/__init__.py b/pkg/suggestion/v1alpha2/__init__.py
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/pkg/suggestion/v1alpha2/parameter.py b/pkg/suggestion/v1alpha2/parameter.py
new file mode 100644
index 00000000000..8918fd877de
--- /dev/null
+++ b/pkg/suggestion/v1alpha2/parameter.py
@@ -0,0 +1,52 @@
+import numpy as np
+from sklearn.preprocessing import MinMaxScaler
+
+class ParameterConfig:
+    """
+    Class to hold the parameter configuration for an experiment.
+
+    Attributes:
+        name_ids (dict): Mapping from a parameter name to the index of that
+            parameter in the other fields.
+        dim (int): Dimension of the vectors created when parameter assignments
+            are mapped to a vector. Each int, double, or discrete parameter
+            adds one to the dimension, and each categorical parameter adds
+            the number of feasible values for that parameter due to one-hot
+            encoding.
+        lower_bounds (ndarray): The lower bounds for each parameter in the
+            search space.
+        upper_bounds (ndarray): The lower bounds for each parameter in the
+            search space.
+        parameter_types (list): The type of each parameter.
+        names (list): The name of each parameter.
+        discrete_info (list): A list of dicts where each dict contains the
+            information for a single discrete parameter. An example of a dict
+            is {"name": "discrete_parameter, "values": [2, 3, 5]}]
+        categorical_info (list): A list of dicts where each dict contains the
+            information for a single categorical parameter. An example dict is
+            {"name": "cat_param", "values": ["true", "false"], "number": 2}.
+    """
+
+    def __init__(self, name_ids, dim, lower_bounds, upper_bounds,
+                 parameter_types, names, discrete_info, categorical_info):
+        self.name_ids = name_ids
+        self.dim = dim
+        self.lower_bounds = np.array(lower_bounds).reshape((1, dim))
+        self.upper_bounds = np.array(upper_bounds).reshape((1, dim))
+        self.parameter_types = parameter_types
+        self.names = names
+        self.discrete_info = discrete_info
+        self.categorical_info = categorical_info
+        if len(self.names) != len(set(self.names)):
+            raise Exception("Parameter names are not unique.")
+
+    def create_scaler(self):
+        search_space = np.append(self.lower_bounds, self.upper_bounds, axis=0)
+        scaler = MinMaxScaler()
+        scaler.fit(search_space)
+        return scaler
+
+    def random_sample(self):
+        new_sample = np.random.uniform(self.lower_bounds, self.upper_bounds,
+                                       size=(1, self.dim))
+        return new_sample
diff --git a/pkg/suggestion/v1alpha2/parsing_util.py b/pkg/suggestion/v1alpha2/parsing_util.py
new file mode 100644
index 00000000000..89566051e5b
--- /dev/null
+++ b/pkg/suggestion/v1alpha2/parsing_util.py
@@ -0,0 +1,144 @@
+"""
+Module containing helper functions to translate objects that come
+to/from the grpc API into the format accepted/returned by the different
+suggestion generation algorithms.
+"""
+from collections.abc import Iterable
+from pkg.api.v1alpha2.python import api_pb2
+import numpy as np
+from .parameter import ParameterConfig
+
+
+def _deal_with_discrete(feasible_values, current_value):
+    """ function to embed the current values to the feasible discrete space"""
+    diff = np.subtract(feasible_values, current_value)
+    diff = np.absolute(diff)
+    return feasible_values[np.argmin(diff)]
+
+
+def _deal_with_categorical(feasible_values, one_hot_values):
+    """ function to do the one hot encoding of the categorical values """
+    index = np.argmax(one_hot_values)
+    return feasible_values[int(index)]
+
+
+def parse_parameter_configs(parameter_configs):
+    name_ids = {}
+    dim = 0
+    lower_bounds = []
+    upper_bounds = []
+    parameter_types = []
+    names = []
+    discrete_info = []
+    categorical_info = []
+    for param_idx, param in enumerate(parameter_configs):
+        name_ids[param.name] = param_idx
+        parameter_types.append(param.parameter_type)
+        names.append(param.name)
+        if param.parameter_type == api_pb2.DOUBLE:
+            new_lower = float(param.feasible_space.min)
+            new_upper = float(param.feasible_space.max)
+        elif param.parameter_type == api_pb2.INT:
+            new_lower = int(param.feasible_space.min)
+            new_upper = int(param.feasible_space.max)
+        elif param.parameter_type == api_pb2.DISCRETE:
+            discrete_values = [int(x) for x in param.feasible_space.list]
+            new_lower = min(discrete_values)
+            new_upper = max(discrete_values)
+            discrete_info.append(
+                {"name": param.name, "values": discrete_values})
+        elif param.parameter_type == api_pb2.CATEGORICAL:
+            num_feasible = len(param.feasible_space.list)
+            new_lower = [0 for _ in range(num_feasible)]
+            new_upper = [1 for _ in range(num_feasible)]
+            categorical_info.append({
+                "name": param.name,
+                "values": param.feasible_space.list,
+                "number": num_feasible,
+            })
+        if isinstance(new_lower, Iterable):  # handles categorical parameters
+            lower_bounds.extend(new_lower)
+            upper_bounds.extend(new_upper)
+            dim += len(new_lower)
+        else:  # handles ints, doubles, and discrete parameters
+            lower_bounds.append(new_lower)
+            upper_bounds.append(new_upper)
+            dim += 1
+    parsed_config = ParameterConfig(name_ids,
+                                    dim,
+                                    lower_bounds,
+                                    upper_bounds,
+                                    parameter_types,
+                                    names,
+                                    discrete_info,
+                                    categorical_info)
+    return parsed_config
+
+
+def parse_previous_observations(parameters_list, dim, name_id, types, categorical_info):
+    parsed_X = np.zeros(shape=(len(parameters_list), dim))
+    for row_idx, parameters in enumerate(parameters_list):
+        offset = 0
+        for p in parameters:
+            map_id = name_id[p.name]
+            if types[map_id] in [api_pb2.DOUBLE, api_pb2.INT,
+                                 api_pb2.DISCRETE]:
+                parsed_X[row_idx, offset] = float(p.value)
+                offset += 1
+            elif types[map_id] == api_pb2.CATEGORICAL:
+                for ci in categorical_info:
+                    if ci["name"] == p.name:
+                        value_num = ci["values"].index(p.value)
+                        parsed_X[row_idx, offset + value_num] = 1
+                        offset += ci["number"]
+    return parsed_X
+
+
+def parse_metric(y_train, goal):
+    """
+    Parse the metric to the dictionary
+    """
+    y_array = np.array(y_train, dtype=np.float64)
+    if goal == api_pb2.MINIMIZE:
+        y_array *= -1
+    return y_array
+
+
+def parse_x_next_vector(x_next, param_types, param_names, discrete_info, categorical_info):
+    """ parse the next suggestion to the proper format """
+    counter = 0
+    result = []
+    if isinstance(x_next, np.ndarray):
+        x_next = x_next.squeeze(axis=0)
+    for par_type, par_name in zip(param_types, param_names):
+        if par_type == api_pb2.INT:
+            value = int(round(x_next[counter], 0))
+            counter = counter + 1
+        elif par_type == api_pb2.DOUBLE:
+            value = float(x_next[counter])
+            counter = counter + 1
+        elif par_type == api_pb2.DISCRETE:
+            for param in discrete_info:
+                if param["name"] == par_name:
+                    value = _deal_with_discrete(param["values"],
+                                                x_next[counter])
+                    counter = counter + 1
+                    break
+        elif par_type == api_pb2.CATEGORICAL:
+            for param in categorical_info:
+                if param["name"] == par_name:
+                    value = _deal_with_categorical(
+                        feasible_values=param["values"],
+                        one_hot_values=x_next[counter:counter + param["number"]],
+                    )
+                    counter = counter + param["number"]
+                    break
+        result.append({"name": par_name, "value": value, "type": par_type})
+    return result
+
+
+def parse_x_next_tuple(x_next, param_types, param_names):
+    result = []
+    for value, param_type, param_name in zip(x_next, param_types, param_names):
+        result.append({"name": param_name, "type": param_type, "value": str(value)})
+    return result
diff --git a/pkg/suggestion/v1alpha2/random_service.py b/pkg/suggestion/v1alpha2/random_service.py
new file mode 100644
index 00000000000..fa31c18d93b
--- /dev/null
+++ b/pkg/suggestion/v1alpha2/random_service.py
@@ -0,0 +1,44 @@
+from logging import getLogger, StreamHandler, INFO, DEBUG
+from pkg.api.v1alpha2.python import api_pb2
+from pkg.api.v1alpha2.python import api_pb2_grpc
+import grpc
+from . import parsing_util
+
+class RandomService(api_pb2_grpc.SuggestionServicer):
+    def __init__(self):
+        self.manager_addr = "katib-manager"
+        self.manager_port = 6789
+
+    def _get_experiment(self, name):
+        channel = grpc.beta.implementations.insecure_channel(self.manager_addr, self.manager_port)
+        with api_pb2.beta_create_Manager_stub(channel) as client:
+            exp = client.GetExperiment(api_pb2.GetExperimentRequest(experiment_name=name), 10)
+            return exp.experiment
+
+    def GetSuggestions(self, request, context):
+        """
+        Main function to provide suggestion.
+        """
+        experiment = self._get_experiment(request.experiment_name)
+        parameter_config = parsing_util.parse_parameter_configs(
+            experiment.spec.parameter_specs.parameters)
+        trial_specs = []
+        for _ in range(request.request_number):
+            sample = parameter_config.random_sample()
+            suggestion = parsing_util.parse_x_next_vector(sample,
+                                                           parameter_config.parameter_types,
+                                                           parameter_config.names,
+                                                           parameter_config.discrete_info,
+                                                           parameter_config.categorical_info)
+            trial_spec = api_pb2.TrialSpec()
+            trial_spec.experiment_name = request.experiment_name
+            for param in suggestion:
+                trial_spec.parameter_assignments.assignments.add(name=param['name'],
+                                                                 value=str(param['value']))
+            trial_specs.append(trial_spec)
+
+        reply = api_pb2.GetSuggestionsReply()
+        for trial_spec in trial_specs:
+            reply.trials.add(spec=trial_spec)
+
+        return reply
diff --git a/scripts/v1alpha2/build.sh b/scripts/v1alpha2/build.sh
index b635ee27f73..caa5f12aada 100755
--- a/scripts/v1alpha2/build.sh
+++ b/scripts/v1alpha2/build.sh
@@ -30,3 +30,6 @@ docker build -t ${PREFIX}/v1alpha2/katib-controller -f ${CMD_PREFIX}/katib-contr
 docker build -t ${PREFIX}/v1alpha2/katib-manager -f ${CMD_PREFIX}/manager/v1alpha2/Dockerfile .
 docker build -t ${PREFIX}/v1alpha2/metrics-collector -f ${CMD_PREFIX}/metricscollector/v1alpha2/Dockerfile .
 docker build -t ${PREFIX}/v1alpha2/tfevent-metrics-collector -f ${CMD_PREFIX}/tfevent-metricscollector/v1alpha2/Dockerfile .
+
+echo "Building suggestion images..."
+docker build -t ${PREFIX}/v1alpha2/suggestion-random -f ${CMD_PREFIX}/suggestion/random/v1alpha2/Dockerfile .
diff --git a/scripts/v1alpha2/deploy.sh b/scripts/v1alpha2/deploy.sh
index 7618edea997..3495191e97f 100755
--- a/scripts/v1alpha2/deploy.sh
+++ b/scripts/v1alpha2/deploy.sh
@@ -38,4 +38,6 @@ kubectl apply -f manifests/v1alpha2/katib/pv/pvc.yaml
 kubectl apply -f manifests/v1alpha2/katib/db/secret.yaml
 kubectl apply -f manifests/v1alpha2/katib/db/deployment.yaml
 kubectl apply -f manifests/v1alpha2/katib/db/service.yaml
+kubectl apply -f manifests/v1alpha2/katib/suggestion/random/deployment.yaml
+kubectl apply -f manifests/v1alpha2/katib/suggestion/random/service.yaml
 cd - > /dev/null
diff --git a/test/scripts/v1alpha2/build-suggestion-random.sh b/test/scripts/v1alpha2/build-suggestion-random.sh
index e1b06019ff0..0277cf81680 100755
--- a/test/scripts/v1alpha2/build-suggestion-random.sh
+++ b/test/scripts/v1alpha2/build-suggestion-random.sh
@@ -39,6 +39,6 @@ cp -r vendor ${GO_DIR}/vendor
 
 cd ${GO_DIR}
 
-#cp cmd/suggestion/random/Dockerfile .
-#gcloud builds submit . --tag=${REGISTRY}/${REPO_NAME}/suggestion-random:${VERSION} --project=${PROJECT}
-#gcloud container images add-tag --quiet ${REGISTRY}/${REPO_NAME}/suggestion-random:${VERSION} ${REGISTRY}/${REPO_NAME}/suggestion-random:latest --verbosity=info
+cp cmd/suggestion/random/v1alpha2/Dockerfile .
+gcloud builds submit . --tag=${REGISTRY}/${REPO_NAME}/v1alpha2/suggestion-random:${VERSION} --project=${PROJECT}
+gcloud container images add-tag --quiet ${REGISTRY}/${REPO_NAME}/v1alpha2/suggestion-random:${VERSION} ${REGISTRY}/${REPO_NAME}/suggestion-random:latest --verbosity=info