Another summer, another edition of our internship on computer vision to be proud of! This time we received well over 100 applications from more than 20 cities including Kyiv, Kharkiv, Lviv, Dnipro, Odesa, Mykolaiv, Vinnytsia, Uzhhorod, Poltava, Kremenchuk, Sumy, Zaporizhzhia, Kryvyi Pih, and Mariupol. What an impressive geography! Only three of the applicants made it to the ‘finals’. Curious what projects they worked on under the mentorship of It-Jim’s engineers? Let’s find out!
The Fifth Edition of It-Jim’s Internships
But first, let’s look at some more numbers. After pre-screening the list of candidates, we reached out to 75 of them and asked them to complete a couple of test assignments. Although only 25 participants sent in their solutions, 15 did so well that they made it to the next step: a technical interview with our engineers. This stage is always a little harsh: we’ve come such a long way together, yet the number of places is always limited and the majority of the candidates, unfortunately, will not receive positive answers. Three of the participants have eventually become our summer interns, and another one even became our trainee (and later a junior CV engineer, but that is a whole different story).
So what were the computer vision tasks our interns were working on for 4 weeks?
Project Zoo
One of the reasons for interviewing prospective interns is to understand their strengths and weaknesses and subsequently provide them with a project that is doable, a little challenging, but certainly educational and broadening their skills.
This summer’s list of projects included:
- Soccer video analytics: creating an app to help assess a soccer player’s agility during practice by tracking the player and the ball and counting the number of kicks the player takes during drills.
- Liveliness detection system: creating a solution that can detect that the facial verification system is being cheated by showing a photo of a person instead of a live face.
- Traffic statistics estimation: creating an algorithm that counts the number of cars and pedestrians crossing a certain line on the road.
Interns’ Solutions
Soccer video analytics
 |  |  |
Demo of automatic kick counting
- Tools and technologies: OpenCV, deep learning, TensorFlow Lite, Kotlin
Liveliness detection system
 |  |
Demo of a liveliness detection system
- Tools and technologies: feature crafting, deep learning
Traffic statistics estimation
Demo of traffic statistics estimation
- Tools and technologies: OpenCV, image processing, object tracking
Summary
Our interns say this program is one of the best ways to get commercial experience and try your hand at being a computer vision engineer. If you’re still wondering if this is the right path for you, remember that you can always try it first. For example, by joining us next winter 2022 for a computer vision internship. We are looking forward to receiving your application!