LLM Inference Engineer
Summary
Architect and maintain high-traffic LLM serving systems, optimizing throughput and latency using inference engines like SGLang, vLLM, and TensorRT alongside GPU programming tools like CUDA and PyTorch.
Locations: San Francisco or Remote
About The Role
The NEAR AI team is building decentralized and confidential machine learning infrastructure to enable user-owned AI. Our mission is to build highly scalable and efficient infrastructure for open-source AI at a global scale.
We are specifically seeking an expert in high-performance LLM serving systems and inference optimization. In this role, you will push the boundaries of how large language models are served.
What You'll Be Doing
- Architect and maintain production high-traffic LLM serving systems.
- Optimize throughput, latency, and cost for leading open-source LLMs.
What We're Looking For
- Strong hands-on experience in LLM inference, with expertise debugging and optimizing major inference engines such as SGLang, vLLM, or TensorRT.
- Deep knowledge of state-of-the-art GPU architectures, and effectively exploit them using PyTorch, Triton, CuTe, CUDA, etc.
- Proven track record in designing and maintaining end-to-end high-traffic LLM serving systems.
- Strong problem-solving skills and ability to communicate technical ideas clearly.
We'd Love If You Have
- Experience with Trusted Execution Environments (TEE).
- Active contributor to open-source LLM inference engines.
Please let us know if you require any special requirements for your interview and we'll do our best to accommodate.