
Working Student – Machine Learning, Computer Vision
Posted 4 days ago

Posted 4 days ago
This is a fully remote position, open to applicants in Germany.
• Conduct research and prototype comparisons of current methodologies in few-shot recognition and visual similarity search.
• Explore DINOv2, SigLIP, and efficient models like MobileNet.
• Utilize publicly accessible food datasets for pretraining and benchmarking purposes.
• Collaborate to create a well-documented dataset utilizing images from actual cafeterias.
• Devise a technique for incorporating new dishes from 1–5 reference images without the need for full retraining.
• Recognize unknown dishes through open-set recognition methodologies.
• Enhance robustness against variations in portion sizes, lighting conditions, and presentation styles.
• Consider domain shifts that may occur between different cafeteria settings.
• Integrate few-shot recognition with detection or segmentation techniques to individually identify multiple dishes on a tray.
• Investigate the potential improvements in recognition through daily menu information and SigLIP’s image–text capabilities.
• Establish metrics and test scenarios while analyzing confusion cases involving similar dishes.
• Develop a camera-based food recognition system tailored for contract catering.
• Currently enrolled in a degree program in Computer Science, Data Science, Mathematics, or a related discipline.
• Engaged in personal projects utilizing Python, whether as part of academic studies, privately, or on GitHub.
• Basic understanding of machine learning and deep learning principles.
• A keen interest in the field of computer vision.
• Experience with PyTorch is considered advantageous.
• Initial exposure to image classification, embeddings, visual similarity search, or object detection is beneficial, though not mandatory.
• Basic knowledge or familiarity with few-shot learning, transfer learning, open-set recognition, or segmentation is preferred.
• Openness to exploring new methods and learning about emerging topics.
• Application materials must be submitted as a single PDF, including a cover letter, résumé, and any relevant certificates.
• Flexible working hours and the option for remote work.
• Choice between part-time and full-time employment.
• Complimentary lunch provided three times a week.
• An open work culture characterized by short decision-making processes and agile methodologies.
• Ample opportunities to make a significant impact.
• Modern office environments located in Hanover or Plauen.
• Fresh fruit, high-quality coffee or tea, and a variety of beverages tailored to each location.
• Regular team events and outings to foster collaboration.
• Bike leasing options available.
• Access to corporate benefits.
• Company pension plan and financial assistance for occupational retirement provision.
• A wide range of professional development and training opportunities.
• Opportunities to attend international conferences.
• Top-floor lounge featuring a rooftop terrace.
Memory Bridge
Entrada
Find Your Talent (FYT)
Inmetrics
Get handpicked remote jobs straight to your inbox weekly.