SpaceTech- Image Processing and Machine Learning Developer

About Vimotek

Space is on the path to industrialization and a new economy is emerging where resources and energy will be extracted, and both products and services will be produced for consumption either in space or on Earth. The first step in this development is for satellites to transition from being standalone passive nodes in an infrastructure to being robots that actively collaborate to solve more advanced tasks. Satellites will inspect and transport each other, refuel, clean up debris, perform repairs, and collaborate to construct larger structures like solar power plants for efficient production of large amounts of fossil-free energy.

Vimotek is a SpaceTech company in Luleå that develops technical solutions which are key components to enable the emerging industrialization of space. The company's first product is targeted at companies that are constructing new satellites to perform various types of advanced missions. The product—the Vimotek NavCam-1 Rendezvous and Proximity Operations (RPO) Kit—is a sensor package that gives satellites the ability to see and understand different objects in space, allowing them to approach and interact with them. Vimotek was founded in 2022 in Luleå.

Vimotek’s solution

The Vimotek NavCam-1 RPO Kit is a subsystem for satellites—a plug-and-play product that will offer spacecraft manufacturers a scalable full solution for spacecraft vision and relative navigation, for use cases in the domains of Rendezvous and Proximity Operations (RPO), On-Orbit Servicing, Assembly, and Manufacturing (OSAM), In-Space Operations and Services (ISOS), Space Situational Awareness (SSA) and Space Asset Protection (SAP).

The product combines three complementary vision sensors (visual camera, thermal camera, and LiDAR) with edge computing power and image processing software to provide actionable navigation data in real-time and ensure successful missions. All sensor systems and associated software are developed in-house by Vimotek.

Our unique approach to sensing and perception in space

The current way for spacecraft manufacturers and operators to gain RPO capability is to build a new technical solution for each satellite and mission, something that is both extremely expensive and does not work for all use cases due to the existing sensor technology not having the right performance. Vimotek’s solution is to develop new high-performance sensors and package them together with intelligent software into a plug-and-play product that can be used by many different customers for many different satellites and cutting-edge use cases. This approach is not the usual way of doing things; Vimotek’s solution is unique in the world.

What have we achieved so far?

Vimotek’s first solution is still under development, but the company has already started to make its mark on the space industry. Most notably, in competition with some of the large market-leading companies, Vimotek’s consortium won the call from the European Defence Fund (EDF) to create a concept for how European satellites can defend themselves against man-made threats in space.

"Bodyguard aims to develop building blocks for autonomous space situational awareness capabilities onboard satellites to increase Europe's independence and superiority in space. The project emerges as a ground-breaking collaborative initiative aimed at elevating the on-board threat detection, assessment and monitoring, as well as protection and resilience capabilities of space assets against a diverse range of threats. In close proximity, it can detect weak points of the threat satellite and counteract with a robot or laser." – The European Commission

What’s the challenge?

As an Image Processing and Machine Learning Developer, you will play an important role within the existing SW team in developing the intelligent perception capabilities at the heart of Vimotek's space vision systems. You'll work with design, training, and deployment of image processing and machine learning algorithms optimized for resource-constrained embedded platforms. This includes fusing multi-sensor data (visual, thermal, LiDAR), creating efficient models for object detection, classification, and tracking, and deploying them to low-power, space-grade hardware. This role offers a unique opportunity to push the boundaries of embedded AI for space applications, collaborating closely with system engineers, sensor designers, and international partners in the space industry.

Skills & Requirements

• Experience in image processing, computer vision, and/or machine learning algorithm development.
• Background in object detection, classification, tracking, and sensor fusion for vision systems.
• Demonstrated ability to design, train, and deploy machine learning models tailored for low-compute, real-time image processing.
• Experience working with lightweight or efficient model architectures (e.g. MobileNet, YOLOv5-Nano, SqueezeNet, Tiny-YOLO, EfficientNet-Lite).
• Solid grasp of quantization, pruning, knowledge distillation, and other model optimization techniques.
• Proficiency in Python and C++, and experience with PyTorch, TensorFlow, OpenCV, scikit-learn, and relevant ML/CV toolkits.
• Experience with data analytics, preprocessing, and augmentation pipelines for multi-modal sensor inputs.
• Familiarity with edge AI deployment and efficient runtime frameworks such as TensorRT, ONNX Runtime, or TensorFlow Lite (Micro).
• Ability to work in a startup environment, balancing innovation with limited resources and high technical ambition.
• Excellent communicator and team player—able to bridge the gap between concept and working product, across disciplines.

Good to have would also be:

• Experience working with space, defence, robotics, automotive, or aerospace domains.
• Hands-on knowledge of thermal imaging, LiDAR point cloud processing, or multi-sensor alignment and fusion.
• Familiarity with deploying models on embedded platforms, edge devices, or using model optimization tools (e.g. TensorRT, ONNX, quantization).
• Experience with FPGA-based image processing pipelines or hardware-aware model design.
• Knowledge of MLOps practices for lifecycle management of ML assets in safety-critical systems.

The company’s primary language is English, proficiency in additional languages is a plus.

Location: Primarily on site in Luleå, but Stockholm or hybrid can be discussed.

Join Vimotek and be part of an exciting journey as we develop cutting-edge products and contribute to industrialization of space.