In addition to pushing the boundaries of existing deep learning technologies, you will be solving complex algorithmic challenges that deep learning cannot. You will be working with a world-class team of engineers to deploy a new wave of AI products that work out-of-the-box across domains without weeks or months of data collection.
- Writing a machine learning pipeline from scratch based on a paper released on Arxiv. This project included developing the dataset loader, augmentation functions, deep learning model, and loss functions — the whole nine yards. And that was just a Tuesday.
- Developing a new probabilistic model to track critical work in manufacturing facilities while ignoring day-to-day variations
- Constraining the problem space creatively in order to develop a practical solution that can be deployed across facilities with very little fine tuning
- Working on novel labeling methods with the data infrastructure team to reduce labeling costs while improving model outputs
- Implementing new loss functions or model architectures so cool they would make waves at CVPR, when you deign to eventually share them
- Using the temporal and consistent nature of video to reduce false positives in a deep learning model while exploring what’s possible in unsupervised object detection
- Have a research and problem solving background given complex problems
- Have built or worked on deep learning training pipelines
- Have worked on robotics and computer vision algorithms (ex: frame transformations, camera calibration, optical flow, etc.)
- Have the ability to quickly hit the ground running and build training pipelines
- Be passionate about AI and is ready to take risks in a fast-paced startup environment
- Competitive compensation and equity packages
- Health, vision and dental insurance (100% coverage for employees, 50% for dependents)
- Flexible vacation schedule