Associate Research Engineer
GKN Aerospace Sweden AB · Full-time
As a Research Engineer specializing in Computer Vision at GKN Aerospace, I led the development of advanced computer vision solutions that significantly enhanced our inspection and quality assurance processes. My role involved designing and implementing both deep learning models using PyTorch and classical computer vision techniques with OpenCV to solve complex industrial challenges.
I was responsible for the complete machine learning pipeline, from initial data collection and preprocessing to model deployment and performance monitoring. This included working with diverse imaging modalities, including high-resolution X-ray scans and standard RGB camera feeds, to develop robust solutions for aerospace component inspection.
My work directly contributed to improving manufacturing efficiency and quality control by implementing cutting-edge computer vision algorithms that automated previously manual inspection processes, reducing human error and increasing throughput.
Key Responsibilities & Achievements
Deep Learning Development
- Training pipeline
- Runtime image augmentations
- Performance/test evaluation
- Ensemble learning
- Segformer & transformer based vision models
- TensorRT optimization
- ML flow logging
- U-Net
- Ultralytics yolov8
- Huggingface
- DICOM data
Computer Vision
- OpenCV
- Scikit-image
- Image processing
- Hough transform
- Algorithm hyperparameter optimization
Data Management
- Data curating
- Data refinement
- Model assisted annotations
- Data annotation conversions
- Label Studio
- Auto-distill
Tools & Workflow
- Microsoft 365 (Teams, Outlook, Office)
- Git version control & GitHub/GitLab
- Sprint meeting format
- Cross-team methodology
- CI/CD pipelines