Stephanie Gredell : Life & Tech

Knowing something is being able to teach it.

A couple weeks ago, I gave a presentation on deep learning two weeks after reading the first part of Deep Learning Illustrated by Jon Krohn.

My company provides a subscription to O’Reilly so I was able to read that book for free through the O’Reilly app. I was searching for a book that I could present a talk to my team. I wanted to give a talk to increase my visibility at work, show my interest in tech outside of my normal, everyday work and just learn something new. While browsing for a book to read, I saw the cover of this book. I think to myself, “Deep Learning? I’ve heard about that. I’ll read this book and give a talk on it.”

Keep in mind, at the time, I had zero interest in Deep Learning. I heard about it from a coworker a few years ago who also did a presentation. However, I did not remember the concepts or what was talked about. I just knew that this thing called deep learning existed. I read the first part of this book and I was hooked. I told a couple coworkers my plan to give this talk. I was met with encouragement from one coworker and skepticism from another. That didn’t matter. It was more about finding the right time to do it.

A week later, my team held a staff meeting and we talked about learning opportunities and exploring new tech. The result of that conversation was to have a regular meeting to show and tell about things we’ve learned about on our own. In my mind, I was thinking “Yes! This is it. This is meant to be. This is when I have to do it!” The meeting was scheduled a week later.

I prepared slides for my talk to discuss the basic concepts of deep learning and then knowing that I’d be presenting to my fellow engineers, I knew I had to present code so there would be something tangible to grasp. There wasn’t really a question about it. However, the coding exercises in the book didn’t really start until part 2 (which keep in mind, I only read part one up until that point). So while polishing my slides, I read the first chapter of part 2 which showed how to code a shallow neural network and worked through the coding exercises on my Kaggle account. I learned of Kaggle from a Youtube video I was watching on supervised learning and thought it’d be perfect since I could link to it.

The talk went over really well. Someone had asked me how long have I been learning about it and they were shocked when I said “Well I only reading part 1 of the book about 2 weeks ago and I coded this up in the last couple days for this meeting.” The whole thing seemed to have gone over really well. I got a lot of direct messages on slack praising me and saying it was really impressive.

Here are the slides and the notes I wrote for my talk for your viewing pleasure.


Leave a Reply

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