We are looking for a Computer Vision Intern to assist in building and refining our image recognition pipeline. The role will start with dataset management—image collection, annotation validation, dataset cleaning, and preprocessing. Once the foundational data work is complete, you’ll get hands-on exposure to model training, augmentation, and evaluation, contributing directly to our production-ready pipeline.
For one in the Seat:
Responsibilities
1.Organize, clean, and preprocess large-scale retail image datasets.
2.Validate and manage annotations (bounding boxes, class labels, segmentation masks if applicable) using tools like Roboflow or CVAT or LabelImg.
3.Apply augmentation techniques and prepare datasets for training.
4.Support in training YOLOv5/YOLOv8-based models on custom datasets.
5.Run model evaluations (Precision, Recall, F1 Score, SKU-level accuracy).
6.Collaborate with the product team to improve real-world inference quality.
7.Document the dataset pipeline and share insights for improving data quality.
Who we're looking for:
Must Have:
1.Basic understanding of Computer Vision concepts (Object Detection, Classification)
2.Familiarity with Python (OpenCV, Pandas, NumPy)
3.Knowledge of image annotation tools (Roboflow, LabelImg, CVAT, etc.)
4.Ability to manage and organise large datasets
Good to have:
1.Experience with YOLOv5 or YOLOv8 (Training, Inference, Fine-tuning)
2.Exposure to image augmentation techniques (Albumentations, etc.)
3.Understanding of retail/commercial shelf datasets or product detection problems
4.Previous internship or project experience in computer vision is a plus