Soma Capital Portfolio Jobs

Software Engineer - Image processing

Destinus

Destinus

Software Engineering
Madrid, Spain
Posted on Friday, April 12, 2024
Destinus is at the forefront of European aerospace, and defence technologies, developing AI-driven autonomous flight systems and advanced turbine engines. Our cutting-edge technologies deliver superior efficiency and performance, pushing the boundaries of supersonic and hypersonic flight. We're pioneering the next wave of combat-proven defence innovations and breakthrough propulsion systems combining turbojet and ramjet technologies for future hypersonic vehicles with hydrogen propulsion.

You will support the team in developing image recognition neural networks. Will have to develop the software for training image synthesis and analysis, based on various backgrounds and images of the typical objects to be identified. These images will then be used to coordinate the UAV flight with surrounding objects at short distances.

Tasks

  • Collaborate with the team to develop and implement image processing software with neural networks suitable for onboard computers.
  • Assist the team in selecting and sampling backgrounds and isolated object images. Pre-process and manipulate images to ensure consistency and quality for training purposes.
  • Generate synthetic training images utilizing various backgrounds and images of the object to be identified.
  • Develop and debug the algorithms for moving objects coordinates definition over the different backgrounds.
  • Debug the onboard software using mostly synthetic data, and then evaluate the performance against real tests.
  • Optimize neural network architectures and hyperparameters to improve performance and efficiency.
  • Evaluate, validate and adjust the model performance through testing and analysis.
  • Stay updated with the latest advancements in image recognition and machine learning techniques.


Requirements

  • Bachelor's degree in Computer Science, Engineering, Mathematics, or related field. Master's degree or PhD is considered a plus.
  • Proven experience in developing and implementing image processing and recognition neural networks using frameworks such as TensorFlow, PyTorch, or Keras and other software for the platforms like RaspberyPi 4 and/or Jetson.
  • Proficiency in programming languages such as C++ and/or Python, with experience in data manipulation and analysis libraries (e.g., NumPy, Pandas). Familiarity with version control systems (e.g., Git) and collaborative development workflows.
  • Familiarity with image processing techniques and tools for pre-processing and manipulating images. Experience with effective image processing libraries like OpenCV and V4L2.
  • Experience with generating synthetic training data and data preparation techniques.
  • Strong understanding of machine learning concepts and techniques, particularly in the areas of convolutional neural networks (CNNs) and deep learning.
  • Experience with generating synthetic training data and data augmentation techniques.
  • Familiarity with image processing techniques and tools for pre-processing and manipulating images.
  • Excellent problem-solving skills and the ability to work effectively in a team environment. Self-motivated results-oriented work is assumed.
  • Strong communication skills (in English), including the ability to explain complex technical concepts to the engineers from the different areas of competence.
  • Demonstrated ability to stay updated with the latest research and developments in image recognition and machine learning.


Destinus is an equal opportunity employer: we value diversity in our companies and our teams. We do not discriminate based on age, race, religion, colour, country of origin, gender, sexual orientation, or disability status