Research Engineer

About The RoleWe are looking for a Video Generation & Editing Model Engineer to join the Pika AI team. In this role, you will focus on post-training and task-specific fine-tuning of cutting-edge video generation and editing models, helping us push the limits of what AI can do in real-world creative scenarios.You will work closely with researchers, product designers, and engineers to build models that power high-quality effects, character replacement, style transformations, and more - ensuring they work seamlessly in production.We are particularly interested in candidates who are passionate about training models for specific functional use cases, and who enjoy experimenting with post-training techniques (such as LoRA, DreamBooth, RLHF, and others) to bring delightful, creator-friendly features to life.What You'll DoResponsible for fine-tuning video generation and editing models, leveraging methods such as LoRA, DreamBooth, and full-parameter fine-tuning in special effects scenarios. Optimize model performance in applications such as effects compositing, character replacement, and style transfer.Design and build large-scale data pipelines and automated annotation workflows to support continuous improvement of model performance in special effects applications.Research and develop fine-tuning strategies and deployment solutions for video generation models (such as Stable Video Diffusion, AnimateDiff, etc.) across diverse application scenarios.Explore and implement best practices for model compression, acceleration, and deployment to enhance real-time video processing efficiency.Collaborate closely with product and design teams to translate cutting-edge algorithms into creator-friendly product features, empowering tens of millions of users.What We're Looking ForStrong passion for technology and a desire to influence the real world through algorithms and tools; highly proactive, self-driven, and able to independently solve complex problems.Excellent engineering practices; proficient in PyTorch, with end-to-end experience covering data processing, model training, and deployment.Practical experience with LoRA, fine-tuning, DreamBooth, or similar techniques in video/image domains; strong understanding of special effects application scenarios.Familiarity with video generation and visual models (such as ViT, Diffusion, ControlNet, StyleGAN, etc.), and a solid understanding of visual generation model training and deployment pipelines.Solid coding skills and strong grasp of algorithm fundamentals; experience with distributed training tools (such as FSDP, Deepspeed, etc.) is a plus.Experience in large-scale data processing, dataset construction, and automated data cleaning is a plus.Experience with AIGC, video effects, or asset generation products is a strong plus.About PikaAt Pika, we're crafting a future where video creation is seamless, intuitive, and universally accessible. Our vision is set on a world where videos transcend entertainment, serving as a canvas for everyone's unique expression. Harnessing AI's transformative power, we aim to dismantle the technical barriers that have traditionally made video production an exclusive field. It's about building the best technology that upholds the power of creativity-not just within our products but as an integral part of our culture.We're a small team of energetic, smart and curious individuals. We value efficiency, collaboration, and creativity. We seek new team members who bring sharp intellect, a strong work ethic, and the ambition to make an impact. Here, you'll find significant room for career advancement, supported by a culture that nurtures both individual and collective development.