I’m an intern

I learned a ton in the ten weeks I spent there–and not only about Machine Learning, which I’ll get to in a moment. My main motivation for pursuing the internship was a desire to further my knowledge of data science and gain a better understanding of what it’s like to work full time in this industry. Now that the ten weeks is up, I can wholeheartedly confirm that I managed to accomplish both goals.

Prior to the internship, I had taught myself most of what I knew by reading, doing online courses and working with simpler data on Kaggle.com. If you add up the amount of time I devoted to my development, it wouldn’t have exceeded 7-10 hours a week. As an intern at deepsense.ai, I had the opportunity to develop 40 hours a week in a very interesting and effective way.

Along with another intern, I was assigned for the entire 10 weeks to a project that one of the teams at deepsense.ai was working on. The project involved developing a predictive modeling model for a client in the advertising distribution industry in applications for mobile devices.

We spent the first week learning about the data we had access to and understanding the business problem to be solved and the technical limitations to deal with as we worked out the solution. The main complications in our project were „Time Series” and „sparsely linked data with user id.”  It would make for good material for the film “Inception 2.”

The second complication was equally inconvenient. Imagine how Netflix recommends movies for you to watch. They base their suggestions mainly on movies you’ve watched in the past, and about other movies that users have watched, who have also watched the same movies as you. In our situation, there was no such feature as a “user” so we could not look at ads that one had seen in the past or ads that people similar to you had seen.

In later weeks we had numerous meetings (at least two per week) with our brain box of a mentor, Mateusz, who taught us a ton. I sincerely thank him for that.

During our meetings we would tell Mateusz how things were going with the jobs he gave us. Once we had talked through a problem, he would give us a new one to solve. It was during these meetings that I learned the most. Beyond updates on the project, we had the opportunity to talk about related topics we didn’t fully understand, analyze our thinking and ask about other concepts we were curious about. Matt was always glad to answer our questions and engage in new discussions. Because he knew so much, our meetings were really educational. They were also among my favorite aspects of the internship.

Beyond the meetings, I learned a lot thanks to just spending time in deepsense.ai’s Warsaw office. Just imagine being in a building where 70 data scientists work, and each of them has more experience than you! (I would really stress that deepsense is bursting at the seams with intelligent people). If I was reading an article on Machine Learning and came across something I didn’t understand and couldn’t get a satisfactory answer to with a quick Google search, all I’d have to do is walk 20 meters to our fun room and there would always be a few people there who could sort it out for me. It was great! We could also ask questions on the company chat, and get the answer within a minute.

Lunches together, a quick game of foosball and video games in the fun room were also a very nice part of the internship. Every day around 12, everyone orders in lunch (compliments of the company, of course). When the food arrived, we’d make our way down to the cafeteria and take a break from work and talk to friends – not always about data science…

A quick game of foosball was also a very popular way to take a break. A special chat channel was even created to organize the quick meetings. All you had to do was put the question out there: “anyone up for a match? We are looking for 2 more people.” 30 seconds later, 4 people would be in the funroom spinning their wrists for ten minutes.

I am very happy with the internship in every respect and plan to return to deepsense.ai next year when I graduate.

I landed on a frontend team, which was creating a UI for Contrail Command. From the get-go, and throughout my internship, all of my team members were extremely helpful and enabled me to quickly come up to speed on the project. During code reviews I learned good practices and improved my overall coding ability. The atmosphere on the team, working with new tools, widening my knowledge with React and learning Typescript were the things that I appreciated the most. And all that under the tutelage of specialists working on a large, commercial project. I’m very satisfied with the internship at deepsense.ai and would highly recommend it to others.

It was a research method that we could use to solve the challenge. With each task I was given, I learned tools and ever more about the project and related data. My buddy and mentor, Marta, helped me clarify all my doubts and solve technical problems. The rest of the team were also all too happy to help me learn. During our daily stand-up, we would discuss the issues we were wrestling with. This kept us up to speed on the status of the work and gave us inspiration to help us solve our tasks. Having lunch together is also an excellent practice. At deepsense.ai, we not only worked together, but also enjoyed integration and other activities which I was thrilled to take part in.

I worked with another intern and together we had the main influence on how we developed the model. Thanks to that autonomy we were able to independently go into numerous issues and look for new solutions ourselves. At the same time, meeting regularly with our team lead, who pointed out the strong and weak points of our choices as well as suggesting other approaches we might try to solve the problem. My three-month internship not only allowed me to develop my data science skills but also to learn how to work better in a team.

They helped me correct my bad habits from my studies and get fully involved in a large-scale project. The benefits were also pretty nice :)

From day one I was treated as a fully functional member of the Hexagon team. This made for a smooth transition into the deepsense.ai workflow. For the first two months I helped my buddy evaluate and enhance models built in of our project. The subequent phases of the project gave me an opportunity to work with 3D point cloud data and graph theory-related issues. During the last few weeks of my internship I was asked to build a demo project using computer vision methods for detecting allergy response on patients’ skin. This was the part that allowed me to learn how to build an end-to-end solution from scratch.

Yet, deepsense is not only about work. It was great that we got to have fun doing the projects, which makes the time fly. We had a few opportunities to have fun outside of work– going to the firinging range, chilling by the river banks and having casual poker evenings in the office. I took part in all of these events and can eagerly confirm that each of them was unique and memorable. In a nutshell, we had great fun every day!