Machine Learning - it's easier than you think!
10 years ago, a younger me interviewed a brilliant young man named Tony for a position as a programmer at Metal Toad. Back then, I didn't know much about hiring (this was long before Metal Toad had implemented the WHO hiring process later recommended by Advisory Board Member, Amanda McPherson) so we were conducting a group interview, lead by me, where we peppered a prospective hire with random questions:
"What's something important about business?" I prompted.
In response, the young programmer went on to speak passionately about ethics in business, and the real burden that technologists carry as it relates to doing right by customers. He had personal experience working at companies where customers had poured out their life savings to build "an app" and either through negligence or willful ignorance, those companies had taken the money and delivered nothing that could have gotten the customers to where they hoped they would be. "Sometimes," he concluded, "it's just as important to tell customers to save their money, as it can be to deliver quality work."
Needless to say, we hired that man on the spot. And over the years, we learned several things about Tony:
- We have yet to find a technology—software or hardware—that he doesn't have an educated opinion about.
- His passion for technology is only matched by his passion for ethical business; he will do the right thing, no matter what.
A year after the interview he was a manager, a year later a director, then a vice president — and today he is our CTO, Tony Rost.
Tony Rost at the AWS Media Solutions Webcast
Metal Toad was honored to speak at the AWS Media Solutions Webcast Series last week. Tony brought his experience and passion to a group of developers and decision-makers eager to get in on the latest developments in metadata and Machine Learning (you can see the entire talk in the video reel below).
Beyond the fascinating nature of the topic — and believe me, Machine Learning is going to change a lot of things — I was incredibly proud of the way Tony spoke. It was not a sales pitch, rather it was a share out of a vision in which the Machine Learning tools at Amazon can be used by all kinds of programmers and companies in ways that simply weren't possible before.
The Quick Version
For those who don't have time for the video here’s a quick highlight reel of the insights Tony presented:
- It's not as hard as it looks. Contrary to popular belief, you don’t need a team of data scientists doing months of R&D to get meaningful results in ML. AWS Media Insights Engine is an entire portfolio of visual machine learning components—and any developer, even junior developers, can launch it in about 10 minutes with a single click.
- You can identify anything. Visual Machine Learning doesn't just do faces. We used AWS Media Insights Engine to bring in ML that adds powerful metadata, like recognition of football teams based on their jerseys. And algorithms can detect products, logos, emotional states, colors, and more.
- The tools are full featured. With the integrated step functions, you can design and run workflows that stitch together services into a single feature-rich application. As an example, we built an online video platform with a suite of traditional features—media repository, basic metadata, relationships to the program guide, etc.
- M&E companies can implement Machine Learning to gather crucial metadata quickly and easily—no data scientists required
- AWS Media Insights Engine allows any developer to create a feature-rich application that will derive business intelligence and improve workflows
- The AWS ecosystem is a huge asset to M&E enterprises by leveraging ML to make the power of metadata easily accessible