We at KTH AI Society strive to bridge the gap between our members and the industry. It’s quite simple, companies working for the solutions of tomorrow need the talent of today. So below we listed all relevant work opportunities for you to take on new challenges.
For more opportunities, visit our strategic partner, Stockholm AI.
If you want to make a job posting contact us at email@example.com.
Any questions? Check out our FAQ
🌍 Scania, Södertälje
Apply before 2022-05-29
Scania is one of the world’s leading manufacturer of trucks and buses for heavy transports, as well as industrial and marine engines. Transport services and logistics services make up an increasing part of our business, which guarantees Scania’s customers cost-efficient transport solutions and high availability. Over a million Scania vehicles are in active use, in over 100 countries.
The goal of this project is to use deep learning methods to generate time-series data that are similar to on-board vehicle sensors (based on collected data) in a decentralized fashion (i.e. with federated learning). Federated learning is an uprising field where we no longer collect data from any entity to the cloud to train a model but instead train a model on many entities and send a trained model back to the cloud. However, there are use cases where having the data on the cloud is better, such as research and developing completely new functionalities. However, collecting data can be GDPR sensitive, as it might be connected to some personal information. AI-generated synthetic data has in recent years become a very popular way to increase data repository, but also a way to use the data for model development or testing functionalities as it is no longer person related. In particular, the field of GANs (generative adversarial networks) has gained huge traction when it comes to generating photos and videos, e.g. generating realistic photos of non-existing persons. Therefore we want to explore this area of decentralized learning of synthetic data generation, such as , , , where we would want to generate realistic time-series data without collecting real data.
The student will have access to Scania’s Hadoop cluster as computational resources to train and run their models.
🌍 Scania, Södertälje
Apply before 2022-06-15
Scania is one of the world’s leading manufacturer of trucks and buses for heavy transports, as well as industrial and marine engines. Transport services and logistics services make up an increasing part of our business, which guarantees Scania’s customers cost-efficient transport solutions and high availability. Over a million Scania vehicles are in active use, in over 100 countries.One basic assumption of multivariate time-series is inherent relationship between the individual variates. In order to exploit the underlying structure in for the purpose of systems modelling, anomaly detection and even in explain-ability of black box AI models, one must learn underlying structures. Therefore, in this thesis, we will build models for structural learning from time-series.
Structural learning is a field that has been of interest for many years . Scania’s interest would be to learn these structural models from observed time-series data. The description of you task is as follows:
Student will be provided access to the computing infrastructure and to the dataset required for the task.
🌍 Backing Minds, Stockholm
Apply before 2022-05-31
We are BackingMinds. We are a Swedish early-stage venture capital company, supporting startup entrepreneurs with backgrounds outside the traditional networks, or founders in overlooked industries. We are driven by our belief that there are exceptional, but overlooked founders or ideas, and we aim to help grow great companies out of that. Today, our portfolio includes companies such as Transfer Galaxy, Dynamic Code and TrusTrace.
We are always looking for driven and talented people eager to work with us within the fast-paced area of startups and entrepreneurs. For the autumn 2022, we are looking for a master thesis student (or a thesis pair) to join our team.
🌍 Babyshop Group, Stockholm
Apply before 2022-05-30
There are hundreds of thousands of users browsing the tens of thousands of articles in our assortment. With such an extensive assortment, finding relevant and inspiring articles becomes far from easy, and the need for recommender systems is high. In ecommerce, providing relevant recommendations is critical for an engaging and satisfactory user experience. Many different methods for creating personalized recommendations exist, e.g., collaborative filtering, content-based filtering and session-based filtering to name but a few. Within the latter, a subset focuses on what is referred to as sequence based recommendations, where the users’ historical behaviour on the site is used to generate recommendations that they would like to interact with in the future.
🌍 SEED EA, Stockholm
Apply before 2022-06-10
SEED is an innovation and applied research hub within EA Worldwide Studios. We bring together passionate team members to explore and build the future of interactive entertainment. Our work combines creativity with state-of-the-art technologies.
Can one use Machine Learning (ML) to drive character body model based on the input speech audio and/or text? What level of control over expressing emotion (e.g. enthusiasm, sadness, happyness, etc.) or other personal trait is possible (stoistic, reserved, friendly, etc.) can we achieve? Is this better than the baseline of selecting between a set of predefined animations?
🌍 Asket, Stockholm
ASKET is a direct-to-consumer clothing brand on a mission to slow down fashion and end over-consumption by creating garments free of compromise, allowing more people to live with fewer but better items. Aside from creating zero compromise garments, a huge part in the success of a modern e-commerce based retailer is to consolidate, manage and act on the data that is being collected from multiple touchpoints.
ASKET has a modern data infrastructure built in order to support automation and data-driven decisions which in short consists of a cloud computing-based data warehouse (DWH) and a customer data platform (CDP). This infrastructure yields datasets which comprise web-interactions, cart creations, transactions and size-related data.
Given these datasets we are now looking to explore how ML can further be used in order to provide best of breed e-commerce. As an example, during last year we developed and launched our proprietary ML-based (SVM) Size Finder with the ambition to increase customer trust and reduce costly and emission driving returns and exchanges. During spring 2022 we are looking for 1-2 students who are looking to explore the latest ML algorithms using highly real and applicable data as their thesis project. Some (not exclusively) of the fields that are of interest are...
Applications are open until November 30th and interviews will be held continuously. Send CV along with some words about what area you think sounds interesting to firstname.lastname@example.org and we'll meet for some coffee and a chat at our HQ at Odenplan.