Engineering Manager, Ray Data

Anyscale is looking for an engineering leader to manage the open source Ray Data team. Ray Data https://docs.ray.io/en/latest/data/data.html is the Python-native data processing engine for AI workloads on Ray. It handles multimodal and structured data and powers batch inference, data loading for distributed training, and large-scale preprocessing. The team owns the open source library end to end, including:

• The streaming execution engine, including scheduling, backpressure, and resource management across CPUs and GPUs
• Batch inference for LLMs and multimodal models, including integrations with inference engines such as vLLM
• Data loading for production training workloads with Ray Train and popular ML frameworks
• Connectors for data sources and formats such as Parquet, Iceberg, Delta Lake, Lance, and cloud object storage
• Performance, stability, and fault tolerance at large scale

This is a first-line engineering management role where you'll lead a team of engineers, stay involved in technical design and code, and own the team's roadmap within the open source project. Ray Data has a large and active community, so you'll also work closely with external contributors to review designs, guide technical direction, and support the project's continued growth and success. The team plays an important role in the broader Ray ecosystem and contributes directly to Anyscale's success.