Machine Learning is an application of Artificial Intelligence. It allows software applications to accurately predicting outcomes in different fields.

Machine Learning is not really new as it has existed for many decades, right from the days of Statistics dominance. Machine Learning in its earlier form was regarded a direct extension of statistics called statistical learning which is the science of learning from large historical data. It is seeing resurgence these days due to the higher information processing and memory processing ability of today’s computing systems.

Why Study It?

Machine learning as a concept is here to stay which will definitely have a lot of applications across all industries as more and more use cases are discovered. As machine learning surge in use cases has been influenced by Big Data the technologies, tools, platforms which make use of Machine learning keeps evolving as big data and the cloud evolves. In all this some hard truth is machine learning will require you to know quite a variety of tools. It is so multi-disciplinary that coming from a financial engineering background I saw myself fit to explore it as my undergraduate involved Statistics and Calculus, Finance and Investment(Accounting and Financial Analysis included) and a cherry on top of Computing courses which included SQL,VBA, Financial Information Systems and programming languages like Java among others. This meant I had the statistics needed to understand result metrics; finance to be able to compare different domain data and get good interpretation of it, and  Java made it easier for me to grasp Python. You cannot be quoted as wrong if you declare Python the Machine Learning language, however R and SQL has maintained a good position in skills needed for a data scientist or machine learning engineer. Machine Learning can be used for pattern recognition, image recognition, speech recognition, predictive analytics etc.  The scope of Machine learning will only keep increasing therefore increasing the demand for professionals able to bring value to various industries using Machine Learning/Deep Learning algorithms.

Some Tips

Master Pandas 

Pandas is a data processing library in Python which helps deal with any form of data in a form similar to that of excel. Pandas is a prominent data-munging tool in Python which uses what we call dataframes(spreadsheet-like)

Understand the essence of Linear Algebra

Numpy is the numerical computing library most ppular in Python


Most machine learning algorithms are available in scikit-learn library. You can do from simple regression, classification, clustering up to more complex NLP use cases.

Understanding Neural Networks

 While DeepLearning may not be a child’s play, to be very effective in Data Science and Machine Learning you have to master tensorflow, pytorch and their related libraries like keras if you are to keep the momentum and be able to tackle any problem thrown at you.

Some Machine learning specialties

• Data Scientists to deploy the algorithms. They often use off the shelf libraries.

• NLP specialists to deal with text including sentiment analysis derived from social media, news and various other sources.

• Data engineers who create Big Data architectures for Machine Learning projects and  ensure system scalability making it more production ready

• Computer Vision engineers usually dealing in image recognition models and various image related projects

• Machine Learning engineers usually used when there is need to deploy some state of the art algorithms like reinforcement learning used in autonomous cars or when there is need to do some research specific for a project.

• Speech recognition engineers if you need some speech recognition or transcribing service done.

Thomas Muserepwa is a Data Analyst/Scientist with JCDecaux. He is passionate about Machine Learning, Data and the Cloud. He can be contacted on

Leave a Reply

Your email address will not be published. Required fields are marked *

Leave a Reply

Your email address will not be published. Required fields are marked *