RDI Solutions

About Rival: Backed by top VCs and angels, Rival is building a unique 3D content-sharing platform and a first-of-its-kind foundational AI model that converts any 2D video into an immersive 3D experience. Currently a team of 13, Rival has brought together talents from Google, Meta, Amazon, BCG, Morgan Stanley, etc. Project Overview: We are seeking a highly motivated PhD intern to join our team and contribute to an exciting project focused on developing a novel, end-to-end system for converting standard 2D videos into compelling 3D (stereoscopic or depth-based) formats using advanced AI techniques. The goal is to research, design, and implement deep learning models capable of understanding scene geometry, motion, and temporal consistency directly from monocular video input to generate high-quality 3D output automatically. This research has the potential to revolutionize content creation and consumption for VR/AR and 3D displays. Your Responsibilities: Conduct literature reviews on state-of-the-art methods in monocular depth estimation, novel view synthesis, video understanding, and 2D-to-3D conversion. Design, implement, and experiment with deep learning architectures (e.g., Transformers, CNNs, GANs, Diffusion Models) for the 2D-to-3D conversion task. Focus on key challenges such as temporal consistency, handling complex motion, maintaining geometric accuracy, and computational efficiency. Process and manage large-scale video datasets for training and evaluation. Collaborate closely with researchers and engineers to integrate findings into a prototype system. Analyze results, document findings, and present progress regularly. Contribute to potential publications or patent applications based on research outcomes. Required Qualifications: Currently enrolled in / just finished a PhD program in Computer Science, Electrical Engineering, Artificial Intelligence, or a related field. Research focus in Computer Vision, Deep Learning, Machine Learning, or Graphics. Solid theoretical understanding and practical experience in deep learning and computer vision fundamentals. Proficiency in Python and deep learning frameworks (e.g., PyTorch, TensorFlow). Experience working with image and/or video data. Strong analytical, problem-solving, and research skills. Excellent communication and collaboration abilities. Preferred Qualifications: Track record of relevant publications in top-tier CV/ML conferences (e.g., CVPR, ICCV, ECCV, NeurIPS, ICML, SIGGRAPH). Experience specifically with monocular depth estimation, stereoscopic vision, view synthesis, video generation, or 3D reconstruction. Familiarity with video processing tools (e.g., OpenCV, FFmpeg). Experience with large-scale model training and data pipelines. Contributions to relevant open-source projects.