Machine Learning Industrial PhD

Publiceringsdatum: 2024-04-22

Är du intresserad av att jobba inom yrken med teknisk inriktning? ChyronHego AB söker just nu efter en dataingenjör och du kan vara den de letar efter. ChyronHego AB behöver någon som vill arbeta heltid. Om du är tillgänglig för 6 månader eller längre, ta en titt på denna jobbannons.

Platsen där jobbet finns ledigt är i Stockholm. Företaget ChyronHego AB erbjuder just nu spännande karriärmöjligheter. Om du vill jobba som dataingenjör har de nu ett ledigt jobb - kolla in annonsen för att se om du har rätt erfarenhet. Bli en del av ChyronHego AB genom att ansöka till jobbet som dataingenjör senast lördag den fjärde maj 2024.

Jobbannons

TRACAB’s Computer Vision and Machine Learning team, based in Stockholm, is looking for an industrial PhD candidate interested in the following project:

Machine learning for discovering patterns in team sports

This project will use state-of-the-art machine learning methods to extract meaningful and tangible information from this data and extend these methods to better serve the domains of human motion capture and sports tracking. Possible approaches include, but are not limited to, graph neural networks, masked auto-encoders and joint embedded predictive encoding, and will utilize semi-supervised, self supervised and unsupervised learning.

This project offers the possibility of analyzing the data at both the team and individual level. On the team level some potential studies include:

• Semi-supervised detection of tactical patterns and events.
• Self-supervised learning for detecting anomalies in multi-agent trajectory data.
• Unsupervised clustering of team constellations with comparable spatial and temporal characteristics. Some hidden patterns that are not hand-crafted by humans can be uncovered by correlating discovered clusters and high-level semantics. This study can contribute to a deeper understanding of complex phenomena in team sports.

The recent release of TRACAB’s Tru-Motion full body player tracking provides a unique new stream of data growing every week, and opens up additional possibilities for detailed study on the player level. This is a less explored research area since we are the first in the world to our knowledge having access to such a large-scale dataset. Potential areas that can be explored include:

• Body-part level modeling of player-player and player-ball interactions.
• Identifying gait signatures and motion characteristics of individual players.
• Identifying and/or predicting potential injuries.
• Assessing a player’s form, emotion and fatigue.

There is also scope to combine team and individual level data.

This position is for 5 years. During this time the student will be employed by the company as an Industrial PhD student (80%) and as a Machine Learning Developer (20%). The position is based in Stockholm.

We are looking for a candidate who has

• A strong desire to work as part of a team and contribute to the company whilst also extending the state-of-the-art and fulfilling the requirements of a PhD.
• An outstanding academic record
• A master in machine learning
• An undergraduate degree in computer science, mathematics, or related field.
• Experience working with soccer data
• Experience with computer vision

If you are interested, please send your application to gareth.loy@tracab.com and include the following:

• A cover letter stating why you are interested in the position
• Your CV
• Your academic transcript