Applied Scientist - Compositor - (Image Restoration)

Key Skills

Computational Photography, Computer Vision (CV), Deep Learning, Image processing, Optimization, Python Programming, PyTorch, Signal processing, TensorFlow

Job Description

Flawless AI is an energetic, growing startup operating at the intersection of AI and film making. Our technology uses generative AI to make it look like your favorite actor or actress is speaking another language, without subtitles or voice-overs, making it indistinguishable from the original performance. It doesn’t matter where you live or what languages you speak, audiences can now experience authentic storytelling exactly as the filmmaker originally intended through the magic of automated visualtranslation. We believe the most important breakthroughs in AI are unlocked only when we apply it to real-world use cases that reach millions of people and are generally available. This is our north star and guiding principle. As an applied scientist on the neural rendering and compositing team at Flawless, you will work with a close-knit, passionate group of world-class researchers tackling some of the most challenging problems in deep learning including multi-modal image synthesis, audio-visual data fusion, and application of some of the most exciting works in generative AI. Our work in automated visual translation is just the beginning, we’re developing countless exciting products based on the application of our proprietary, cornerstoneresearch. This is an unbelievable opportunity to join a team operating at the cutting edge of the generative revolution, don't hesitate, reach out today.

Requirements

• MS or Ph.D. in Visual Computing, Computer Vision, Machine Learning, or related field

• 3+ years of proven industry or research experience with deep learning and computer vision in two or more of the following:Computational Photography (mandatory) e.g. inpainting, blur, grain, color-space based modeling, camera models/optics

• Domain adaptation techniques for image face synthesis/editing/restorationNeural talking head synthesis

• Un-supervised, self-supervised, few/zero-shot learning applied to image synthesis/editing/restoration problems

• 3+ years of experience in Python with proficiency in deep learning frameworks such as PyTorch or Tensorflow

• Experience developing tools and solutions at scale for visual effects or real-world computer vision problems

• Proficiency in signal processing and numerical optimization

Preferred Qualifications

• Experience with 3D GANs for human face synthesis/editingExperience with multi-modal image synthesis/editing and audio-visual data

• Implemented models from top-tier publications in venues such as Siggraph, ICCV, CVPR, ICML, or NeurIPS.

Benefits

• Autonomy - You'll own your work from start to finishInfluence

• You'll impact major research decisionsPublication

• You'll be encouraged to publish work through collaborations with researchersLearning

• You’ll push the state-of-the-art with the best in the worldImpact