Machine Learning Resident - Client : Outsyders

“If you are interested in leveraging Generative AI for Computer Vision in visual effects, film industry, and gaming, this is the right opportunity for you. Be a part of a team of research and machine learning scientists and get mentored by some of the best minds in AI while doing it.” - Kunwar Saaim, Machine Learning Scientist

Description

About the Role

This is a paid Residency that will be undertaken over a twelve-month period with the potential to be hired by our client afterwards (note : at the discretion of the client). The Resident will be reporting to an Amii Scientist and regularly consult with the Client team to share insights and engage in knowledge transfer activities.

About our Client

Outsyders are redefining the art of theatrical 3D conversion. Founded by pioneers from three of the most respected stereoscopic conversion studios, our team fuses decades of experience with cutting-edge machine learning to push the boundaries of cinematic immersion. Our innovations are trusted by the biggest names in Hollywood. We don’t just convert films; we elevate them, delivering a premium 3D experience that enhances storytelling, deepens engagement, and sets a new industry standard.

About the Project

This project focuses on developing generative AI techniques for video inpainting in stereo conversion, leveraging diffusion models and spatiotemporal architectures to reconstruct missing or occluded regions in monocular video. In stereo conversion workflows, generating a second view from a single camera feed often results in occluded regions that need to be inpainted with depth-aware, temporally consistent information.

Unlike traditional methods that rely on heuristic propagation or adversarial learning, this research explores diffusion-based generative models to synthesize high-fidelity missing regions while maintaining structural, temporal, and depth consistency.

Who Are You

You have completed a graduate-level program or higher (M.Sc / Ph.D) in Computing Science, Machine Learning, or Engineering, with substantial research and project experience in deep learning, computer vision, and generative AI. You are proficient in Python and experienced with key ML frameworks and libraries such as PyTorch, Hugging Face Diffusers, vLLM serving and inference, and OpenCV. Your positive attitude towards learning new applied domains and your ability to communicate technical concepts clearly make you a valuable team player. You are enthusiastic about collaborating across interdisciplinary teams to achieve excellence.

What You Will Be Doing

In this role, you will develop GenAI computer vision models specifically tailored for Outsyders’ immersive technology applications to enhance interactive experiences, contributing directly to the advancement of Outsyders’ products and services.

Your responsibilities include preparing and curating relevant datasets, utilizing commercial and open-source models and tools to support GenAI computer vision tasks, and optimizing model pipelines to ensure scalability and real-time processing. You will collaborate with cross-functional teams to develop client-centric solutions, participate in client meetings, and contribute to reports on project progress.

Required Skills / Expertise

We’re looking for a talented and enthusiastic individual with solid knowledge of machine learning, deep learning, and computer vision.

Key Responsibilities

• Design, implement, and optimize diffusion-based generative models for visual effects, mainly focussing on video inpainting and masking, while maintaining structural, temporal, and depth consistency. Refine solutions for industry-specific challenges.
• Prepare, curate, and preprocess high-quality large scale datasets for training or fine-tuning, and validating models, integrating diverse sources such as images, videos, and 3D assets for visual effects and content creation.
• Leverage state-of-the‑art computer vision and GenAI frameworks and tools, including both commercial and open‑source libraries to enhance model performance, accelerate workflows, and optimize data processing.
• Collaborate with cross‑functional teams to build and deploy computer vision‑based minimum viable products (MVPs) that address client needs, ensuring seamless integration into existing systems.
• Engage in regular client meetings, contributing insights and updates on model performance and project milestones through presentations and detailed reports.
• Optimize computer vision pipelines to ensure efficient, scalable, and real‑time inference capabilities, leveraging techniques like model quantization, transfer learning, and distributed processing.

Required Qualifications

Preferred Qualifications

Non-Technical Requirements

Why You Should Apply

Besides gaining industry experience, additional perks include :

#J-18808-Ljbffr