Developer Technology Engineer - AI
Core
Build and optimize core parallel algorithms and data structures for GPUs to solve performance bottlenecks in large language model (LLM) training and inference, contributing to frameworks like Megatron, TRTLLM, SGLang, and vLLM.
Role type
Senior IC GPU software engineer (LLM optimization)
Builds
Production GPU kernels, communication libraries, and open-source frameworks for distributed LLM workloads
Domain
Artificial Intelligence / High-Performance Computing / GPU Architecture
Deliverable
production ML models
Required skills
C/C++/Python/Fortran, parallel programming, GPU programming, algorithm design, distributed systems, compiler optimization, linear algebra, numerical methods
Preferred skills
LLM framework optimization, distributed communication optimization, NVLink/InfiniBand/RoCE protocols, open-source contribution
Technologies
Megatron, TRTLLM, SGLang, vLLM, cuDNN, cuBLAS, CUTLASS, DeepGEMM, FlashMLA, FlashAttention, Flashinfer, NCCL, NVSHMEM, DeepEP
Responsibilities
Optimize GPU kernels and instruction-level tuning for LLM workloads; Refine distributed training and inference communication libraries; Collaborate with architecture and research teams to influence next-generation software platforms; Engage in deep optimization of high-performance operators and compiler settings
Seniority
Senior, hands-on IC
