At .txt, the business problem is to make AI reliable by developing software that goes beyond simple conversation, specifically through structured generation. The company is building products like dotjson that guarantee JSON Schema formatted LLM output, and they need a senior engineer to contribute to these efforts.
Requirements
- Professional experience with Python and Rust
- You have designed and implemented distributed systems making use of asynchronous network programming
- Experience with GPU programming using pytorch
- experience writing CUDA kernels
- experience with LLM inference servers like vLLM, SGLang, TensorRT-LLM
Responsibilities
- Build and maintain efficient, reliable Python and Rust code for .txt products like dotjson
- Collaborate with compiler and research engineers to integrate the backend services with our methods for structured generation.
- Build and improve systems for understanding the performance of our code at every level, from our low level Rust libraries to our integrations with modern LLM inference servers
- Maintain high code quality and enforce best practices through code reviews, testing, and a penchant for meaningful commit messages.
Other
- You know how to use version control for collaborative workflows with git on GitHub.
- You are comfortable frequently sharing unfinished work through draft pull requests and are committed to making your work understandable to your teammates.
- Significant public work that we can use to demonstrate role-relevant expertise.
- Please provide a 1-page resume in English.
- Kindly, do not apply to more than one position at a time.
