Remote · Junior · up to 200,000 ₽/month
Responsibilities:
- Data collection and labeling
- Training computer vision models for detecting violations in images: (garbage/pollution, infrastructure damage, non-compliance with standards)
- Using OCR models (Tesseract, PaddleOCR) with post-processing, normalization, and data validation
- Developing AI chats using RAG (retrieval-augmented generation), designing prompt engineering and chain logic
- Analyzing geotrails: identifying deviations, matching photos, coordinates, and time, detecting anomalies
- Working with locally deployed ML/AI models
- Participating in the design of solution architectures and algorithm selection
- Consulting on specialized topics
Requirements:
- Experience with Python and frameworks PyTorch / TensorFlow / Keras;
- Practical experience using OCR systems (Tesseract, PaddleOCR, EasyOCR);
- Experience in image processing (OpenCV, PIL, torchvision);
- Knowledge of convolutional network architectures (ResNet, EfficientNet, ConvNeXt, YOLO, DETR, etc.)
