AI Compiler Research Intern (MLIR, GPU)

ByteDance; San Jose (CA); On-site; USD 100,000 - 167,000; Part time

Job summary

ByteDance invites an intern to join the Seed Infrastructures team in San Jose to contribute to AI compiler optimizations for training and inference workloads. You will help develop MLIR-based passes, optimize model execution on GPUs and NPUs, and support deployment pipelines, while collaborating with researchers to translate requirements into compiler and runtime improvements.

The 12‑week internship offers hands‑on experience with distributed training, large-scale ML systems, and

Qualifications

• Pursuing a PhD in CS/EE or related field.
• ),

Responsibilities

• Contribute to AI compiler optimizations for training and inference workloads.
• Develop and extend MLIR-based compiler passes for graph lowering, optimization, and code generation.
• Optimize model execution on GPU and NPU accelerators, focusing on performance and memory efficiency.
• Support model deployment pipelines, including compilation, packaging, and runtime integration.
• Assist with distributed training and inference acceleration, such as parallel execution and runtime scheduling.
• Benchmark, profile, and analyze performance across hardware backends.
• Collaborate with researchers to translate model and system requirements into compiler/runtime improvements.

Skills

Open source LLM frameworks experience

DL framework familiarity

Computing systems understanding

Compiler/model optimization knowledge

Distributed ML systems experience

GPU/TPU/NPU programming

AI compiler stacks knowledge

12-week internship eligibility

Education

PhD candidate in CS/EE or related

Job description

ByteDance invites an intern to join the Seed Infrastructures team in San Jose to contribute to AI compiler optimizations for training and inference workloads. You will help develop MLIR-based passes, optimize model execution on GPUs and NPUs, and support deployment pipelines, while collaborating with researchers to translate requirements into compiler and runtime improvements.

The 12‑week internship offers hands‑on experience with distributed training, large-scale ML systems, and