R&D Engineer, Training Infrastructure

Our team provide foundational infrastructure and solutions for ML/AI scenarios at ByteDance, including data management for machine learning samples and multimodal data through Magnus, as well as distributed orchestration and computing framework capabilities through Primus and Ray.

• Participate in the design and development of Ray-based distributed computing solutions within the company, supporting the Seed large model business.
• Contribute to feature development and performance optimization for the Ray, KubeRay, and related upper-layer ecosystem frameworks.
• Build capabilities for running Ray on Kubernetes, including elasticity, tidal resource clusters, stability, observability, and platform integration.
• Contribute to the Ray open-source community.

• Bachelor’s degree or above, preferably in Computer Science or a related field.
• Proficient in programming languages such as Python and C++.
• Experience with the RayCore/RayData/RayServe or Ray-related frameworks.

• Familiarity with common distributed computing frameworks, such as Spark/Flink; machine learning background is a plus.