Machine Learning Engineer at Winnow

Cluj-Napoca, Romania

true · full_time · Visa sponsorship

Apply for this role at Winnow

Responsibilities:
- Design, prototype and evaluate computer-vision and multimodal ML models to detect and classify food items from images/video.
- Build and maintain training pipelines, data augmentation, annotation workflows and evaluation suites.
- Optimize models for deployment on edge hardware and cloud (latency, memory, accuracy); produce ONNX/TensorRT builds where needed.
- Integrate models into production stacks, working with firmware/edge/cloud teams to deploy and monitor inference pipelines.
- Implement CI/CD for model training and deployment, and set up experiment tracking and performance monitoring.
- Collaborate with product, data and ops teams to prioritize features, label datasets, and iterate on real-world failure modes.

Requirements:
- Master’s degree (or equivalent) in Machine Learning, Computer Science, or related field; PhD is a plus.
- Proven end-to-end experience developing and shipping ML models (training, validation, deployment) in production.
- Strong Python skills and deep learning framework experience (PyTorch/TensorFlow); familiarity with transformers and vision models.
- Experience with model optimization tools (ONNX, TensorRT) and edge platforms (e.g., NVIDIA Jetson) and cloud (AWS).
- Solid knowledge of data annotation practices, dataset design, and evaluation metrics for object detection/image classification.
- Good communication skills and ability to work cross-functionally in agile teams.