Machine Learning Systems Engineer, Research Tools

Design, develop, and maintain tokenization and encoding systems used across Pretraining and Finetuning workflows; optimize encoding techniques to improve model training efficiency; collaborate with research teams to address data representation needs; build infrastructure to enable experimentation with novel tokenization approaches; implement monitoring and debugging for tokenization in the training pipeline; create robust testing frameworks for tokenization across languages and data types; identify bottlenecks in data processing related to tokenization; document systems and communicate decisions across teams. Strong software engineering experience with machine learning, proficiency in Python, and experience with ML data pipelines or infrastructure are valued. Familiarity with tokenization algorithms (e.g., BPE, WordPiece) and distributed ML workloads is a plus. Must be eligible to work in the US; Anthropic sponsors visas for some roles. Deadline to apply: rolling; Location-based hybrid policy requires 25% in-office presence in listed offices (New York, NY and San Francisco, CA). This role sits within the Encodings and Tokenization team and supports research progress toward reliable, interpretable, and steerable AI.