Role overview
Join our world-class computer vision team as we revolutionize horticulture automation. Whether you're a recent graduate eager to make your mark or an experienced engineer looking for your next challenge, you'll develop cutting-edge perception systems that enable robots to understand and interact with complex greenhouse environments - from identifying ripe produce to detecting plant diseases and optimizing crop health.
As a Computer Vision Engineer at Eternal, you'll be part of a high-performance culture that values first-principles thinking and rapid iteration. You'll work at the intersection of classical computer vision and modern deep learning, creating perception systems that operate reliably in challenging agricultural environments with varying lighting, occlusions, and organic variability.
You'll collaborate with a distributed team across our Cologne HQ and Bengaluru office, pushing the boundaries of what's possible in agricultural computer vision while delivering practical solutions that work 24/7 in production environments.
Responsibilities
- Design and implement robust computer vision algorithms for crop detection, ripeness assessment, and precise localization in dynamic greenhouse environments
- Develop deep learning models for multi-class segmentation, object detection, and tracking of plants, fruits, and agricultural structures
- Create real-time perception pipelines that process 2D/3D sensor data for robotic decision-making with sub-centimeter accuracy
- Build intelligent systems that adapt to varying environmental conditions, including changes in lighting, plant growth stages, and seasonal variations
- Optimize vision algorithms for edge deployment on robotic platforms, balancing accuracy with computational efficiency
- Implement continuous learning systems that improve model performance through data collected from our deployed robot fleet
- Collaborate cross-functionally with robotics engineers, AI/ML researchers, and crop scientists to deliver end-to-end perception solutions
- Bachelor's degree in Computer Science, Electrical Engineering, Applied Mathematics, or related field (or graduating by Summer 2025)
- Strong programming skills in C++ and/or Python for computer vision applications
- Understanding of fundamental computer vision concepts: image processing, feature detection, camera calibration, and 3D geometry
- Experience with deep learning frameworks (PyTorch, TensorFlow) and classical CV libraries (OpenCV)
- Familiarity with Linux environments and version control systems
- Passion for solving complex real-world problems with tangible impact
- Recent graduate or final year student with strong academic performance
- Hands-on computer vision experience through internships, research projects, or competitions
- Demonstrated programming skills through coursework or personal projects
- Understanding of CNNs and basic deep learning architectures
- Solid foundation in both classical and deep learning-based computer vision
- Experience deploying at least one vision system from research to production
- Proficiency with modern architectures (YOLO, Mask R-CNN, Vision Transformers)
- Understanding of model optimization techniques and edge deployment
- Proven track record of deploying vision systems in production environments
- Experience with 3D vision, multi-sensor fusion, or SLAM algorithms
- Knowledge of model optimization for embedded systems (quantization, pruning, distillation)
- Ability to mentor junior engineers and lead technical initiatives
- Technical leadership experience with complex perception systems
- Deep expertise across multiple vision domains (2D/3D, classical/learning-based)
- Strategic thinking about perception architecture and technology roadmaps
- Track record of building and scaling high-performance computer vision teams
Preferred qualifications
- Experience with agricultural or outdoor computer vision applications
- Knowledge of 3D sensors (stereo cameras, LiDAR, structured light)
- GPU programming skills (CUDA) for accelerating vision algorithms
- Experience with vision-language models or foundation models
- Familiarity with ROS2 for perception system integration
- Publications at top-tier computer vision conferences (CVPR, ICCV, ECCV)
- Open source contributions to computer vision projects
- Vision Libraries: OpenCV, PCL, Open3D
- Deep Learning: PyTorch, TensorFlow, ONNX
- Deployment: TensorRT, OpenVINO, ONNX Runtime
- Sensors: RGB cameras, stereo vision, depth sensors
- Infrastructure: Cloud-native training pipelines, edge deployment systems
- Integration: ROS2 for robotic system integration