top of page

Hackmakers wraps up the #FormulaAi Hackathon

Updated: Jun 17, 2022


The winning team of the hackathon have won a round trip to the Red Bull Racing Factory in Milton Keynes.🏆🏆


All the finalists as well as the honourable mentions and newcomers also won a suite of fantastic prizes worth over $30,000 AUD!



18th March 2022, Sydney


The FormulaAI hackathon kicked off on 18th of February 2022. The first of its kind with a theme centred around “Formula 1 Racing ” in collaboration with Oracle. The four days-long virtual hackathon was amazing, with 2000+ participants from more than 30+ countries. Over 130 mentors, who had expertise in a number of fields, helped the teams throughout each phase of the hackathon event.


The themes for the hackathon were in the fields of Data Analytics, Augmented Reality and 3D Modelling. The problem statements for each of the 3 challenges were as follows:

  1. Data Analytics: Create an accurate weather prediction model for the Formula 1 2021 video game.

  2. Augmented Reality: Build the Next-Gen swift UI/UX for the upcoming metaverse AR app.

  3. 3D Modelling: Create a blender python script that translates JSON word coordinates into a beautiful 3D race track.

There were more than 140 teams who submitted brilliant ideas and solutions for solving the problems. Based on their video presentations, accompanying codebase, and any additional documentation, their work underwent a review from a panel of judges. Their solutions were judged by the following criteria: consistency, originality, quality and design, and commercialisation opportunity the winners have been selected.


There are 3 teams total that secured the 1st, 2nd and 3rd positions in the hackathon. We have also awarded 3 teams with honourable mentions and 3 new teams for their impeccable performance and their solutions.




The Winners


1st Position : Team DolphinGang

Led by Marina Chau, Team DolphinGang is from France and are from the University of Paris. The team worked on a forecast prediction model that predicts the weather and the rain percentage at every corner of an F1 track. They also improved the model by adding new features such as the wind, the humidity, and more!.



2nd Position : MIT Boys

Led by Harihaaran B, team MIT Boys is from India, and they are students at Manipal Institute of Technology. They have set out to programmatically create a 3D-Model of an F1 track that would be scalable to most tracks currently on the F1 calendar. This has culminated in a ready-to-use Blender Python script that renders a 3D Model, utilising geometric coordinates of the track data.



3rd Position : Team Kaha

Led by Austin Chamberlain, team Kaha is a combination of participants from New Zealand, Australia, India and Pakistan. Named Project RERE, team Kaha worked to create a model to make weather predictions based on observations and historical data. They also created an app to combine the model output with other data to present an information dashboard for race teams.





Honourable Mentions


Please note that the numbering and position in the below list is not a ranking, as the Honourable Mentions are not ranked.



Honourable Mention 1 : Team 57B

Led by Zohaib Qazi, Team 57B is from Australia and includes professionals who are currently working at the Commonwealth Bank of Australia. Together, they formed a solution that aims to arm Oracle Red Bull Racing Esports with intelligence on weather events at strategic intervals, in order to better inform decision-making on the track in real time. The model will automatically tell teams what localised conditions to expect at 5, 10, 15, 30 and 60 minutes ahead of time.



Honourable Mention 2 : ARce

Led by John Trujillo, Team ARce participated in challenge 2. The team had participants from Ecuador, Pakistan and the United States. Together they brought Data Storytelling to life for Formula AI. This took the form of a tool for sharing your data in a comfortable and immersive way. Some of the features of their app include rotate and zoom, distance measurement, and general information.




Honourable mention 3 : Fantastic 5

Team Fantastic 5 was led by Yamini Harikrishnan with team participants coming from Canada and India. They participated in challenge 3. Their solution is a Blender add-on that allows anyone to generate an F1 circuit track without the need to manually model anything. The add-on comes with a simple-to-use interface that presents the user with the option to select the track that needs to be modelled, alongside the particular sectors that need to be included in the 3D model. It sends a request to an API which then downloads the track information on behalf of the user.

Moreover, props on the track such as light towers and flags can be easily configured using the scrollers. The cost of adding more features in the future is dramatically reduced since everything is open-sourced.

The process of modelling an F1 circuit is fully automated, with the end result being generated in a matter of seconds; this process can take hours to make by hand.





Newcomers Award


We also have awarded prizes for newcomers into the hackathon. These newcomers have given a significant amount of effort and have provided a good solution for the problem statement.


Newbie 1 : Team Kuku

With Ege Hosgungor as the leader, this team participated in challenge 1. The participants were from the United Kingdom and Turkey, graduates from the Koç University. Team Kuku formed a custom ML pipeline for both classification and regression models for the data provided. They have wrapped the system with docker and deployment is done on the cloud.




Newbie 2 : Team TwinAI

With Zion Pibowei as the leader, team TwinAI participated in challenge 1. The participants included professionals from Games Consort and students from the University of Lagos. Coming from Nigeria, the team made a product named TwinAI. This product is an engine for weather forecasting for Formula 1 racing. The system consists of two models stacked together, one for predicting weather type and the other for predicting rain percentage probability. The models take in input data containing specific information from a user, preprocess the data to carry out the classification and regression tasks, then outputs inference results as a JSON response.



Which were the projects you liked?

Let us know by commenting below !!

73 views0 comments
bottom of page