Often the first steps in AI are the hardest and too many good ideas will never become reality… though we are convinced that it doesn’t need to be that way! We created an intense AI hackathon for students and hackers to show that you can get from an idea to a live running AI prototype in just 14 days! Execution is king, but it all starts with the right people and ideas. We teach our students:

  • fall in love with the problem and not the solution in order to stay focused but
  • to test their hypothesis as quickly as possible based on real customer feedback and iterate
  • to set clear goals, deadlines and execute
  • to have fun

The winning teams presented their projects in front of more than 200 people at the Demo Day at the Google office in Munich and we were excited to see what they came up with!

 

Fast shopping with ‘Pickeru’

The audience chose team ‘Pickeru’ to win the ‘Think Award’. Their AI solution fastens the inconvenient queuing and checkout process in brick and mortar stores. Users can clip their smartphone onto a small bracket at the shopping basket and use the Pickeru app to automatically scan items placed inside the basket. Their computer vision techniques run on the device and showed high accuracy. The team is eager to continue with their idea and to push it to the next level.

 

Calculation of production time with ‘Predict.it’

The ‘Make Award’ for the best technical implementation went to the team ‘predict.it’. The two students Florian and Muhammad developed a software that gets CAD files of small technical parts as input and then uses a neural network to predict how much time e.g. a milling machine would need to produce it. This reduces expensive manual calculations and increases throughput. Florian’s grandfather, who was in the audience as well, will definitely be among the first customers.

 

Reading emotions with ‘EmoConn’

The ‘Start Award’ for the best startup potential was handed over by our hosts from Google and went to the team ‘EmoConn’. They developed an emotion analytics platform for YouTube influencers that provides insight about how viewers respond to which part of the content. Through face recognition techniques and head pose estimation, the system can tell how engaged the users are and suggests changes to increase the quality of the video clips. Not only OpenMind, a YouTuber from Munich, showed strong interest in their product but also a large influencer marketing platform that might be among their first B2B customers.

 

About the organizer

At appliedAI we work with top AI startups, early-stage founders and AI hackers to help them have an impact. If you are interested in contributing or participating, click here to reach out to us.