Student Researcher - Compiler - 2026 Start
Core
Develop AI compiler optimizations for training and inference workloads, focusing on graph lowering, optimization, and code generation using MLIR.
Role type
Researcher, AI Compiler Engineer
Builds
Optimized model execution on GPU and NPU accelerators, distributed training/inference pipelines
Domain
Artificial Intelligence, High-Performance Computing, Compiler Technology
Deliverable
production ML models
Required skills
MLIR-based compiler passes, GPU/NPU optimization, distributed training/inference acceleration, model profiling and benchmarking, open-source LLM inference frameworks (vLLM, SGLang), deep learning frameworks (PyTorch, Megatron, DeepSpeed, JAX), modern computing systems architecture
Preferred skills
Large-scale ML systems optimization (FSDP, DeepSpeed, Megatron, GSPMD), HPC and communication technologies (CUDA, Triton, NCCL, RDMA), AI compiler stacks (torch.fx, PyTorch Dynamo, XLA)
Technologies
MLIR, vLLM, SGLang, PyTorch, Megatron, DeepSpeed, JAX, CUDA, Triton, NCCL, RDMA, torch.fx, XLA
Responsibilities
Develop and extend MLIR-based compiler passes for graph lowering and code generation; Optimize model execution on GPU and NPU accelerators; Support model deployment pipelines through compilation and runtime integration; Benchmark and analyze large-scale models across hardware backends; Collaborate to translate model requirements into compiler improvements
Seniority
Researcher, Academic/Student
