Software Engineer, Tokens and Prompt Structures
Core
Design and maintain encoding libraries and prompt structures that enable Claude to consume text and multimodal data, bridging research experiments to production systems.
Role type
Senior Software Engineer (LLM Infrastructure & Encoding)
Builds
Core encoding libraries, SDKs, and data abstractions used across Anthropic's pretraining, finetuning, and API systems.
Domain
Artificial Intelligence / Large Language Models / Data Encoding
Deliverable
production ML models | product features
Required skills
Software engineering, library/SDK/API design, complex refactoring, ML/LLM architecture understanding, performance optimization
Preferred skills
Tokenizer experience, long-term library maintenance, Python, Rust, reinforcement learning infrastructure
Technologies
Python, Rust
Responsibilities
Maintain and improve encoding libraries used by engineers and researchers; Run experiments to optimize structured data feeding into Claude; Design data structures and abstractions to shield the organization from encoding details; Adapt libraries to support new research directions and ship to production; Optimize encoding performance across dependent systems
Seniority
Senior, hands-on IC