Fake Your Way to Technical Wizardry with Michael Lee
“Thought I would try to build a rough proof of concept using computer vision…”
The words popped across my screen, just as I was attempting to muscle my way through writer’s block. It was a perfectly timed interruption. I popped over to a group message thread that I’m on with some of my friends and former colleagues from Care.com to learn that Michael Lee had built an app that provides feedback on your posture while biking.
The crazy thing is that this isn’t even the first app that Mike built. Mike studied Economics as an undergrad and got a Masters degree in Interactive Media. He’s never worked as an Engineer, and yet he’s now built two proof of concept mobile apps with machine learning and computer vision. If Mike can do it, so can you.
As PMs, we have many ideas, but we often worry that we’re not technical enough or don’t know how to get started. We may not always know enough technical details to talk to engineers about it, and we’re scared to admit it. New technologies are exciting, but can be daunting. I sat down with Mike to learn more about how he grew his technical prowess.
Here are Mike’s five strategies on how PMs can be fluent with new technologies, even if you’re feeling a little technically timid.
#1: If you want to be more technical, doing a side project is a great way to learn
Doing a side project isn’t the only way to learn, and it depends on how each person learns best. However, for Mike, getting hands-on, end-to-end experience cements concepts much better than reading about it or playing with an app that uses the technology. It has also helped him develop greater empathy for the teams that he works with. That alone is worth the time spent doing a side project. Getting hands-on experience is a great way to learn — not just about new technologies, but any skill that you want to build.
#2: Build something that you’re interested in
Mike loves biking, so that passion is what inspired his last two projects. His recent project was an app that analyzes your side body position while riding a stationary bike. After reading Model Zoo, Medium, and LinkedIn Feed posts, he realized that Google ML Kit had recently released some software that would make it possible for him to hack together an app. On the software side, he found code for an existing camera app on GitHub and used XCode to integrate Google ML Kit into the app code. On the hardware side, he used two phones. One phone captures the side position and does pose estimation, and the other phone sits in front and mirrors the display so users can see their side position without turning their head.
His first project was an app where you could take a picture of a bike gear and it would tell you whether the gear was worn out or not. To do that, he had to painstakingly create a data set by collecting pictures of bike gears, setting up bounding boxes on each image, and labeling each tooth of the gear on each image. It took about 3 days to install all the packages onto his computer, 5 days to create a data set of 100 images, and 3 days to train a new model with his data. Building something that you’re interested in makes your effort not just about learning, but also engaging with your passion.
#3: Don’t worry about taking a windy path when your goal is to learn
When your goal is to learn, don’t worry about stumbling here or there. Sometimes, you’ll do things the long way, because it’s something you want to learn about. Mike admits that he learned a ton about data labelling in the five days that he spent assembling the training data set. He initially set the bounding boxes incorrectly and had to go back and re-do many images. This makes him much more aware of how having well labeled training data is a great advantage for tech companies.
Other times, you should take shortcuts, because it’s not important to what you want to learn. For example, Mike’s goal wasn’t to practice writing front end app code, so he found app code that he could tweak or pare down to what he needed and kept the UX in the app spartan with only a splash screen and camera screen. He also didn’t want to spend a ton of time figuring out how to stream from one phone to the other, so he spent $1.99 buying an app in the app store that does just that. Even though the streamcast functionality wasn’t natively embedded into his app, the experience was close enough to enable him to simulate whether a two device setup would work.
#4: Learning can take as much time as you have
Most PMs will have to make do with working on side projects in your free time. For PMs in research heavy roles however, Mike recommends that you set aside more of your time to play with new technology. Regardless of which type of PM you are, it’s critical for you to stay up-to-date on new technologies, and know when to leverage them in your products.
Mike has made the most of his learning time by employing these strategies:
- Map out how much of your free time you want to spend on your side project. Even if you don’t feel like you have a lot of free time, that’s ok. You can fit it into the time that you do have.
- Timebox and set goals for each work session. As you run into problems (which will definitely happen), Stack Overflow is a helpful place to look for answers. Think about each work session as a sprint and evaluate what goes into each sprint before automatically carrying it over.
- If you can’t get something to work within the allotted time, cut your losses and try another approach. For example, if your app code isn’t working, then go back to Github and look for a different app that you can use as your base code.
#5: Share your work to continue the learning
Don’t keep your side project to yourself. When Mike shared a post about his app with his network on LinkedIn, he knew that his network would be supportive, but he was surprised by how helpful the comments were. They gave him new use cases to consider, ideas to improve his app, and other tools to try out. Sharing your work unlocks even more avenues to continue learning, and if you keep learning, you will become more technical.
Like the slogan in Gatorade’s famous campaign from the ’90s, “Be Like Mike.” Conquer any self-doubt that you have about not being technical enough. Be like Mike. Be curious and challenge yourself to learn about new technologies. You don’t have to have a technical background to learn about new technologies. Regardless of what happens with Mike’s apps, he’s already achieved his goal. He’s accelerated his understanding of machine learning and computer vision by building something.
And just in case you’ve been wondering for the last thousand words — no, we’re not related. :)
