Lead development of high-performance Rust parsers, compilers, automata, indexes, and data structures for structured language-model generation. Optimize decoding-loop performance through profiling and benchmarking, integrate Python bindings and C++/CUDA components into inference engines, collaborate across teams, and maintain production-quality code through testing and CI/CD.

Develop high-performance parsers, compilers, and data structures that constrain LLM decoding according to schemas, grammars, policies, and function definitions. Research parser, compiler, and symbolic methods; write reliable, efficient code; and optimize implementations through rigorous testing and automation. The role requires professional Python expertise, lower-level programming experience, strong formal language knowledge, and clear written communication. It is fully remote with at least three hours of daily overlap with Eastern Time business hours.