The TUM.ai Makeathon is a virtual 48-hour challenge where you work on a real world business case and implement an AI application with a team of talented participants from various backgrounds. Every team gets access to a pool of powerful AI models, which stem from different domains, such as, Regression Analysis, Computer Vision, and Natural Language Processing.
Each team submits two deliverables "Business Model Canvas" and "First Prototype with an Open-Source model usage", and each team requires one to two participants with a background in IT. The primary goal of the competition is to create the most innovative business idea with a foundational AI application.
To support you, we will provide helpful workshops explaining the models, as well as assistance during the event from Industry Experts. Fantastic prizes await the winners, one of which is a membership to TUM.ai.
Videos of our last Makeathon!
Click here to view recordings of insightful talks and the final presentation round of our GPT3-Makeathon back in April 2021.
The main idea of our Makeathon is to develop a solution with a cross-functional team, which is based on one of the models from our Open-Source Pool. The solutions you will be working on have to fit into one of four tracks: Finance & Investment, Environmental Sustainability, Infrastructure & Public Services and Social Media Platforms. Within these tracks, you can either join a challenge pitched by one of our industry partners (Challenge-Setter), or if you have an awesome idea yourself you can also pitch it, so you can convince others to join you in building a viable business concept.
Regression analysis is a predictive technique (forecasting) which helps to predict business insights. This approach is also used to detect which attributes directly influence the observed outcome.
Computer Vision is based on the two operations detection and classification. An object can be easily detected, given at least one image as input data, all while simultaneously being classified.
Natural Language Processing gives the ability to computers to understand text and speech, at the same level as humans do. These AI models also enable the verbal communication between human and machine.
Submit your application before the 3rd of October.
You will get informed via email whether you have been accepted and will get to be part of this amazing event.
Here you get the chance to either pitch your own idea or listen to the other pitches. Afterwards, a team finding event will take place where you can discuss the presented projects with the other participants and form you interdisciplinary team. These events will take place in the evening.
Use the workshops to get helpful insights into AI as well as how business modeling works. Join business talks and get inspired by entrepreneurs or industry experts. The workshops and business talks will also happen in the evening and the exact dates will follow.
You have 48 hours to develop a proof of concept with your team in order to provide the following deliverables at the end of the Makeathon:
After pitching your results on Sunday afternoon, a jury of experts will select the winning teams for each track. Fantastic prizes are waiting!
Of course, you don't work all the time,
stay tuned for fun activities including
a meme challenge.
Meet and socialize with many
new people despite the restrictions
imposed by COVID-19.
Win a Mentorship Program
sponsored by the TUM Venture Labs
or other cool prizes.
Get to know students
from other backgrounds and
learn from each other.
Get exclusive access
to the world´s most powerful
language AI to tackle real-world
Join insightful workshops
and business talks and get help
from mentors and experts
from the industry.
The whole concept of our Makeathon is to build upon the idea of cross-functional teams. Therefore, we are looking forward to all kinds of skill sets relevant for our Makeathon (see question "What if I already have a team?"). AI will impact every domain. Yours too. Join and find out how!
A Makeathon is a fun event that brings together like-minded people from different backgrounds to collaborate and rapidly build new innovative products.
No, you don’t need a team. A team-finding event will take place at the beginning of October where you can form your cross-functional team which can consist of up to 5 people.
In order to fulfil the deliverables "Business Model Canvas" and "first prototype with an AI model usage", your team should cover tech, AI, and business skills as well as domain knowledge for your corresponding track (see application formula). As long as this is fulfilled and your team has a max. of 5 people, you can work with whoever you would like. Therefore it will not matter if you already have a team before the start of the event.
Note: Of course, one person can have more than one skill or domain knowledge, for example, two participants could already fulfil the four requirements (tech, AI, and business skill and domain knowledge).
We will provide a jury of experts who will examine your solutions and evaluate them fairly.
Contact us at firstname.lastname@example.org, we will surely find a solution ;)
Everyone around the world is welcomed to apply. After the application deadline on the 30th of September, a selection process will follow since we have a certain limit on the number of participants.
The event will take place online. Apart from your laptop, your daily supplies and some sleep, you don't need anything.
This year the event will take place online. But rest assured, we have planned the day to nonetheless be unforgettable. So it will be a unique opportunity to get to know other students, which is quite hard in times of Corona.
Feel free to contact us at email@example.com.
The main event will take on the weekend from 15th to 17th of October. Here you use the 48 hours to the best of your ability to develop your innovative idea together with your teammates. The workshops, business talks, challenge pitches, and the team finding event will be very casual and happen during the evenings beforehand. The exact dates for these events will be announced at the beginning of October via email.