ML Engineer (Computer Vision - All Levels)

Role overview Develop computer vision models that turn live camera footage into useful decisions for real-world settings. The work spans the model lifecycle, from preparing data and evaluating performance to investigating failures and optimizing models for deployment. The opening welcomes candidates at different career stages, with level matched to experience.

Responsibilities - Train, fine-tune, and evaluate models for detection, segmentation, tracking, and recognition use cases. - Clean and prepare datasets, then create evaluation sets to assess whether models are ready for use. - Export and optimize models for production inference, collaborating with experienced ML engineers. - Write maintainable, tested Python code that teammates can use and extend. - Investigate errors in real camera footage and use findings to improve model behavior. - Share results and lessons through concise technical write-ups and reviews.

Requirements - Strong Python skills and hands-on experience training models with PyTorch. - Working knowledge of computer vision methods, including CNNs or transformers for detection, segmentation, or classification. - Familiarity with OpenCV and model evaluation measures such as precision, recall, and mAP. - Understanding that model performance can change with new data, and ability to investigate those failures. - Comfort working with Linux, Git, and Docker. - Curiosity, ownership, and clear communication; experience may range from recent graduate to established engineer.

Nice to have - Experience with ONNX, TensorRT, video datasets, or vision-language models.

Benefits and work setup The source describes a remote, small-team environment with opportunities to learn from engineers deploying models in production and to take ownership of work.