I've worked with the 3 main cloud providers over the past decade. Whilst AWS has an abundance of services + excessive docs, and Azure is strictly for .NET refugees, for me, GCP (Google Cloud Platform) is the preferred cloud provider - for the simple reason that it has the optimal Tensorflow and Kubernetes management infrastructure. In order to architect ......

---

Graph Theory Conceptslast year I was working on an NLP Deep Learning project that required me to compare parse trees for different question / answer pairs. I needed to extract feature-set for my model, so I leveraged NetworkX to represent my data as comparative graphs. NetworkX https://networkx.github.io/ is a Python package for the creation, manipulation, ......

---

Scikit LearnScikit-learn ( http://scikit-learn.org ) is a free software Machine Learning lib for the Python programming language. It features various classification, regression and clustering algorithms including support vector machines, random forests, gradient boosting, k-means and DBSCAN, and is designed to interoperate with the Python numerical ......

---

basic Machine Learning pre-requisites:python http://docs.python-guide.or... http://jupyter-notebook-beg... http://www.scipy-lectures.o... https://pandas.pydata.org/p... http://scikit-learn.org/sta... charts https://seaborn.pydata.org/... ......

---

OverviewTensorFlow https://www.tensorflow.org/ is an open source software library for Machine Learning across a range of tasks, developed by Google. It is currently used for both research and production at Google products, and released under the Apache 2.0 open source license. TensorFlow can run on multiple CPUs and GPUs (via CUDA) and is available ......

---

Keras QuickRefKeras is a high-level neural networks API, written in Python that runs on top of the Deep Learning framework TensorFlow. In fact, tf.keras will be integrated directly into TensorFlow 1.2 !Here are my API notes:Model APIsummary() get_config() from_config(config) set_weights() set_weights(weights) to_json() to_yaml() save_weights(filepath) ......

---

Big Data keeps evolving. Stone Age Hadoop was alot of Java bolierplate for defining HDFS access, Mapper & Reducer. This was superceded by Bronze Age Spark, which provided a succint Scala unification of:ML pipelinesin-memory structured DataSets over RDDs via a SparkSession SQL APIDistributed Streams(Note: You can run such jobs easily in a dynamically ......

---

The Docker CLI commands actually encapsulate Rest calls to the host Docker Daemonhttps://docs.docker.c... daemon listens on unix:///var/run/docker.sock - which you can curl into: curl --unix-socket /var/run/docker.sock http:/containers/jsonIn theory, a Container app could be Cluster-enabled via startup ......

---

HashiCorp Consul https://www.consul.io/ provides an easy to use, multi-region Service Discovery / Health-Check + distributed config keyval store. If you have ever hacked away at coding ZooKeeper to support distributed systems, Consul's Agent based architecture requires an order of magnitude less effort to roll out.Here's my Consul QuickRef:CLI options:agent ......

---

Redis is an in-memory NoSQL store that can store 5 types of data: Strings, Lists, Sets, Sorted-Sets and HashMaps. It is a good choice for fast reads/writes in distributed systems for a relatively small dataset size that is memory-capacity bound.v2.8.x is supported by AWS Elasticache for a DevOps-free deploy, which supports replication strategy of 1 ......

---