Machine Learning as a Skill Set

Problem

Coding is similar to explicitly writing the recipe for a dish; spelling out every granule of sugar that needs to go into the cake. The reality of that is laborious, and humans simply can’t remember and visualize large amounts of data or even read it without feeling bored.

Actions taken

After 15 years of coding in this way, which can be quite tedious at times, I invested my personal time and curiosity into machine learning. I was convinced that computers learning from examples rather than explicit instructions would be massive and I just had to get a piece of it.

Lessons learned

The amazing thing about science and machine learning is that it can infer and guess, and do things off the script that even humans could not spot across such a large amount of data. As a product leader, I was looking to understand and leverage that so that I could make great products.

When you start to realize how that works, you become inspired by the range of products you can build. As an engineer or product person, you form an appreciation for where that technology can be applied and how it can change human lives.

I think everyone should appreciate how far you can go in a rules-based system versus how far you can go in an example-based system. Once you understand those two things, then you start to understand what can be built and invented.