Have you ever wondered why traditional education might not be everyone’s cup of tea? Well, a theoretical approach simply does not work for everyone. As they say, practice makes perfect, so why not engage our students in the material via problem-solving activities, we thought. And this approach seems to work for our juniors! Through the mole detection project, our juniors got a taste of what a career as an AI data scientist looks like. Curious about the project and how our students experienced it? Be sure to read the article below.

Learning is a process rather than a collection of technical knowledge. In order to survive in this competitive world, knowing how to put your acquired tech skills into practice is so vital. And what better way is there than learning by doing!

Via hands-on student projects, we invite our juniors to put their knowledge to the test and thereby give them a taste of what a job in the field looks like. This was no different for the mole detection project that challenged our juniors to develop and deploy a model that can detect skin cancer. “A company, called Skin Care, requested us to develop a web application where the user can upload a picture of a skin mole and that can, in turn, predict whether this mole is dangerous or not. If the mole has proven to be dangerous, the user should get the advice to go see a doctor”, explains AI junior Sara Silvente. 

In order to start developing the model, the juniors first had to collect the necessary data. “Once we were divided into groups, we started communicating with the client to get some extra clarification and information. The client then provided us with a dataset that contained images of benign and malignant moles”, explains AI junior Naomi Thiru. But before using this dataset to build and train the model, some extra steps had to be taken. “Once you have the dataset, you have to prepare it by cleaning it and expanding its diversity to make the predictions even more accurate. Afterward, you fit your dataset into your model for it to be able to predict whether or not a mole is benign or malignant”, Naomi continues. 

But a big part of the project also focused on deploying the model, in other words, making it available to the users. “You may have succeeded in creating a model, but if it’s not usable, it’s pretty much useless. Therefore the goal of this project was to create a model and make it available to the users. To do so, we used Heroku, a cloud-based service on which you are able to run and deploy such a web application”, explains Naomi. 

Challenging sales pitch
Soon the next challenge was lurking around the corner: presenting their project to the client. “As we had to present our project to different people from the company, including sales and marketing, we had to keep it simple and had to avoid too much jargon. This type of presentation was therefore very different from the presentations that we had given so far. They focused much more on the technicalities, rather than outlining what we’d created”, explains Naomi. 

Sara’s team, on the other hand, dealt with this challenge in a very creative way. “It was tricky to present the project to someone who doesn’t know anything about data science. As we had to keep it easy, we explained how the model worked in a superficial way. We also decided to sell our application as a product, called Mole Care. Talking to your audience from a business perspective made it easier for us to avoid going into details.”

The importance of soft skills
Technical skills, such as knowledge of Python, machine learning algorithms, and computer vision were crucial in bringing this project to a successful conclusion, yet the importance of soft skills can not be underestimated. “Working in a team was really important in this project, but I really love doing it. There is always someone who knows that one thing that you don’t know, that is more advanced in a certain subtopic or that has seen something that you’ve missed out on”, explains Naomi. 

Both Naomi and Sarah emphasize that they have been able to sharpen their soft skills during various projects, rather than having acquired new soft skills during this specific experience. “We’ve worked on several projects during the training, which allowed us to get better at collaborating. We’ve learned how to rely on each other’s skills and how to contribute our own skills to these kinds of projects”, explains Naomi. “Presentational skills, team working, adaptability, etc. are skills that we’ve sharpened project after project. Each project has helped us improve. We’ve learned much more via these projects than during our individual learning path”, Sara continues. And let this be BeCode’s strong point according to her. “An individual learning path is integrated into almost every self-learning course, but what’s interesting here is the group projects. They are more than an added value, it’s the essence of BeCode.

Are you interested in following the AI Bootcamp?

Amazing! – We have a new class starting in Brussels, Ghent, Antwerp, Charleroi and Liège.