diff --git a/README.md b/README.md
index 7d0ec89..48b2d5b 100644
--- a/README.md
+++ b/README.md
@@ -1 +1,24 @@
-# spark-mpi-tf
\ No newline at end of file
+# SPARK-MPI-TF
+
+This project demostrates the Spark-MPI approach within the context of Spark-based TensorFlow distributed deep learning
+applications. The direction is addressed by several other projects, such as
+[BigDL](https://github.com/intel-analytics/BigDL) and
+[TensorFlowOnSpark](https://github.com/yahoo/TensorFlowOnSpark). In comparison with these alternative
+solutions, Spark-MPI aims to derive an application-neutral mechanism based on the MPI Process Management Interface (PMI)
+for the effortless integration of Big Data and HPC ecosystems. 
+
+## Prerequisites
+
+1. [Spark-MPI](https://github.com/SciDriver/spark-mpi): PMI-based approach for integrating together the Spark platform and MPI applications
+
+2. [Horovod](https://github.com/uber/horovod): MPI-based training framework for TensorFlow
+
+## Examples
+
+The MNIST Spark-Horovod [IPython notebook](https://github.com/SciDriver/spark-mpi-tf/blob/master/examples/mnist/spark_horovod.ipynb)  for handwritten digit classification (see, for reference, [TensorFlow Tutorial](https://www.tensorflow.org/tutorials/layers)).
+
+
+
+
+
+
diff --git a/examples/mnist/README.txt b/examples/mnist/README.txt
index 5c999ef..228b51e 100644
--- a/examples/mnist/README.txt
+++ b/examples/mnist/README.txt
@@ -1,6 +1,9 @@
+This Spark-MPI example demonstrates the integration of Horovod's MPI-based deep learning engine
+with the Spark platform within the context of the MNIST application. 
 
-Running Spark-MPI with Jupyter
--------------------------------
+
+Running with Jupyter on a single node
+-------------------------------------
 
 export HYDRA_PROXY_PORT=55555
 
@@ -9,4 +12,3 @@ export PYSPARK_DRIVER_PYTHON_OPTS='notebook --no-browser --port=7777'
 
 pyspark --master local[*]
 
-pkill -9 "hydra_pmi_proxy"