Ashley Sanders
Dec 18, 2017 · 4 min read

It’s never easy to approach and experiment with new technologies right now, where the tech stacks and languages you’re already familiar with are constantly evolving, or being declared deprecated altogether.

I believe the only way you can truly open your mind to new tech is to dive in head first, practically, and apply it to something that matters to you.

Get Going

With that said, here is a relatively simple how-to-get-started reference:

  1. Understand why — Make it clear to yourself why you want to build AI and how you can apply it to the real world.
  2. Choose your weapon — There are many different directions you can take, from beginner tools like Weka Workbench, to more complex and powerful open source libraries like TensorFlow, or even the high level statistical systems like R. Go with the one/s that you feel you’ll be the most productive with, given your current experience in programming and mathematics.
  3. Have a process — Define your problem, prepare your data, check your algorithms and train/test and improve your results.
  4. Practice, practice, practice — You can never be too comfortable with a new tool. Practice with simple in-memory data-sets, practice with real problems, and most importantly practice with problems that matter to you.

Machine Learning Methods

There are an array of machine learning methods, which should be used depending on the required outcome, the two most commonly referred to being Supervised and Unsupervised.

The core differences between the two methods are that Unsupervised learning technically has no right answer and is used predominantly when the required outcome is less structured and more about finding interesting results, whereas Supervised learning methods are trained to give accurate predictions and are more structured. The majority of practical machine learning is done using Supervised learning methods. Here is an in-depth comparison and breakdown of each by Dr. Jason Brownlee. 👈



There are many different algorithms you can use within your models. Mathematics is at the core of ML and prior comfort in algebra particularly will help you understand how everything works — even though libraries like TensorFlow have baked-in activation functions for you to make us of, having a knowledge of their core will be a great advantage. Use of ML libraries is inevitable, but understanding how they work is even more important.

Application and uses

There are an increasing amount of use cases for artificial intelligence, below are some of them in no particular order:

  • Speech recognition
  • Face recognition
  • Data analysis and prediction
  • Artificial creativity
  • Computer vision, virtual reality and image processing
  • Game theory and strategic planning
  • Game artificial intelligence/computer game bots
  • Diagnosis (artificial intelligence)
  • Natural language processing
  • Algorithmic trading

The list can go on and on, in almost all fields of science and otherwise — medicine, finance, human resources, psychology, robotics… there is almost no limit to what you can do with it.


Machine learning is one of the most interesting and powerful topics for engineers today, and is probably one of the most challenging to tackle, but given the right tools (the internet is littered with them) you can really dive right in and surprise yourself with what you can achieve.

At Hi5 we aim to produce amazing data. By using machine learning and data science we’re able to do just that — and gain valuable insights into how companies and people are performing culturally using our platform.

A scatter plot graph showing the correlation between giving and receiving recognition within the workplace using Hi5 — the more you give, the more you receive, and vice versa

A note for early stage development

Think of your model as a brain that needs to learn. Feed it garbage data and it’ll learn garbage, e.g. if you are going to be training your model to be able to tell you what type of bird it sees when presented with an image, there’s no need to feed it unnecessary amounts of cat pictures. Keep it clean and on point. Unfortunately, data preparation can become the most time consuming task of all.

If you found this helpful or interesting in anyway, please give it a 👏 so others can find it, too!

Thanks 🙌

P.s. Interested in seeing what your company’s people data will look like? Why not try Hi5 today?

Getting better, together.

Learnings and thoughts on getting better, together. By the makers (and friends) of Hi5.

Ashley Sanders

Written by

Software Engineer | @giveaHi5

Getting better, together.

Learnings and thoughts on getting better, together. By the makers (and friends) of Hi5.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade