Looking for students to join us in solving manufacturing's biggest challenges using AI.

Looking for students to join us
in solving manufacturing's biggest
challenges using AI.

AI Academy

16.02.2022

Powdered by AI

Who are we?

We are Oqton’s AI team. Our ultimate goal is to fully automate the manufacturing process – from design to production to logistics – using state-of-the-art AI technologies.

We deliver the AI models and the supporting infrastructure that powers Oqton’s Manufacturing Operating System.

We analyze all kinds of data, 3D point clouds, sensor data, audio data, imagery, ... to support applications such as 3D part segmentation, orientation for 3D printing, reverse engineering point cloud scans to CAD models, 2D and 3D nesting, job scheduling, tool wear, machine monitoring…

Reverse Engineering - Converting a 3D point cloud to a CAD representation

What can we offer you?

  • If you like to apply state-of-the-art AI technology while working on your master thesis, you can contact us for a MSc thesis topic to work on.
  • If you want to gain some industrial experience and earn some money during the summer holidays, you can apply for an internship.
  • If you are interested in performing world-class PhD research with industrial relevance and business potential, we can engage in a Baekeland PhD project (With support from VLAIO)

3D Point Cloud Segmentation - Identifying regions of interest within a 3D point cloud

What can you expect?

The challenges we love to work on together with you:

  • Nesting, also known as bin packing, which involves placing a collection of 3D parts in the limited build envelope of the 3D printer (the maximal physical volume the printer can build)
  • Scheduling, which concerns allocating tasks to be executed to available resources such as 3D printers.
  • Segmentation, which covers identifying local regions of interest within a 3D point cloud
  • Reverse engineering, which supports turning a 3D point cloud scan into a CAD representation
  • Data reliability, which covers analyzing the quality of our datasets, consisting of millions of 3D objects, and hence guaranteeing the performance of our AI models
  • Anomaly detection, which examines if a machine’s behavior is not “normal”.

AI technologies we (would like to) use for such applications:

  • Different network architectures, such as rotation-invariant point cloud networks, transformers, general adversarial networks, variational auto-encoders, …
  • Different algorithms, such as advanced search algorithms, genetic algorithms, no-fit polygon…
  • Different learning paradigms, such as reinforcement learning, federated learning, self-supervised learning…

In addition to working on such challenges and technologies in a stimulating R&D environment alongside highly motivated people, Oqton offers:

  • access to large real-world datasets
  • use of our state-of-the-art compute infrastructure
  • mentoring by seasoned AI engineers

If you want to find out more about Oqton's Manufacturing Operating System and how we use point cloud based deep learning for supporting digital manufacturing, have a look at this video.

Watch video

Tom Tourwé

We are also open to consider your interests and to think along with your ideas, contact us with your own creative suggestions!

Tom Tourwé

Head of AI

What do we expect

Our ideal candidate:

  • is curious and eager to explore state-of-the-art AI technologies
  • is able to work autonomously
  • has good programming skills in either Python or C++
  • loves to tackle complex challenges, in our case related to geometric deep learning, combinatorial optimization, or multimodal learning