Python and Clarifai — Hacker’s Delight

Jumping into a new project can be a daunting task at best, or downright scary at worst. I’ve attended a few Hackathons now since starting my software engineering degree in 2014, but even still, without fail, about half way through the weekend I get the sinking feeling in my stomach that there’s not a chance my team will be able to finish even a fraction of our goals; we’ve bitten off more than we can chew; bugs are everywhere; we haven’t slept in what feels like days and there’s no way we’ll be able to figure anything out now!

That’s how I was feeling when I started on my task of implementing Clarifai into our voice activated photo capture and recognition application. Our goal was to have the camera take a photo, then for the photo be described to the user after being analyzed by Clarifai. Given the difficulties we had run into thus far getting all the moving pieces of our tech puzzle to fit together, I wasn’t optimistic when I first delved into the Clarifai API documentation. However, just a few minutes after starting to read the reference, I stopped, confused, and wondered:

“That’s…it? What am I missing? Is there another page or document I’m missing?”

But there wasn’t. That really was it. In 3 or so lines of code I had made my first API request and about an hour later my team had developed an algorithm that filtered the API’s response to our liking. We could capture an image then listen to Clarifai’s description of it only moments later.

Hacking together a new piece of tech in as little as 24 hours will never be a cakewalk, but Clarifai made the bumpy ride just a little bit smoother. After only barely cracking the surface of what’s possible with the robust and easy to use image recognition API, I’m looking forward to having another go at it, and exploring more of the many features it has to offer.