Model Training Data Platform Engineer

About the Role

We are looking for a Model Training Data Platform Engineer to join Pika’s AI team. You will be responsible for building and optimizing data workflows for model training, supporting our cutting-edge video and image generation models. You will collaborate closely with researchers and engineers to ensure large-scale datasets are efficiently managed, processed, and served to power Pika’s product innovations.

This position can be contract-based or full-time, and is open to candidates who are based remotely or in Palo Alto.

Responsibilities

• Drive the construction and optimization of large-scale data pipelines for video and image model training. Take ownership of data collection, cleaning, labeling, and serving processes.

• Build scalable and reliable data platforms from the ground up, enabling researchers to easily access and utilize datasets for model training and fine-tuning.

• Design and implement automated data labeling and quality control workflows to improve the efficiency and accuracy of training data preparation.

• Support various model training tasks, including diffusion models, video generation models, and other visual content generation models.

• Explore innovative uses of LLMs to enhance data labeling, augmentation, and metadata extraction processes.

• Collaborate closely with researchers and product teams to ensure that data pipelines and tooling align with evolving model requirements and product needs.

Requirements

• Bachelor’s degree or above in Computer Science, Software Engineering, or a related technical field.

• Strong coding and system design skills; proficient in Python (preferred), with solid experience in building scalable data platforms and pipelines from scratch.

• Hands-on experience with large-scale data processing, data cleaning, automated data labeling, and data quality management.

• Familiarity with AI-generated content (AIGC) workflows, diffusion models, and visual content generation technologies.

• Experience working with LLMs (Large Language Models) for data-related tasks (such as automated labeling or data augmentation) is a strong plus.

• Proactive, self-driven, and capable of independently solving complex engineering problems. Passionate about building high-impact tools to accelerate AI research.

About Pika

Pika is building the next generation of video creation tools powered by AI. We aim to empower creators around the world with intuitive, powerful, and intelligent tools that unlock new forms of storytelling and creativity. Our team is pushing the frontier of video generation and editing technologies, with a fast-paced, product-driven culture.