Role Overview
Intel is building the next generation of software technologies for heterogeneous computing. The Compiler Engineering Graduate Intern will contribute to an LLVM-based compiler stack for CPUs, GPUs, and accelerators, supporting AI, HPC, and demanding workloads.
Responsibilities
- Contribute to the design, implementation, testing, or analysis of compiler and runtime technologies.
- Develop and debug software in C or C++ within a large, collaborative codebase.
- Create or improve tests, reproduce defects, analyze failures, and help deliver reliable solutions.
- Evaluate functional correctness, runtime performance, or code quality.
- Participate in code reviews, technical discussions, and project planning with mentors and team members.
- Document designs, experiments, results, and lessons learned, and present project outcomes.
- Contribute to open-source projects when aligned with the assigned work.
Requirements
- Currently pursuing a Master or PhD degree in Computer Science, Computer Engineering, Electrical Engineering, Applied Mathematics, or a related technical discipline.
- Graduate-level coursework, research, or project experience in compilers, programming languages, computer architecture, parallel computing, systems software, or a related area.
- Experience programming in C or C++ and working with complex software systems.
- Experience designing experiments, analyzing technical results, and communicating conclusions.
- Familiarity with LLVM, MLIR, Clang, Flang, SYCL/DPC++, Triton, OpenMP, CUDA, or another compiler or parallel-programming ecosystem.
- Hands-on experience with LLVM or related compiler infrastructure, including front-end development, intermediate representations, optimization passes, code generation, runtimes, or debug information.
- Research or development experience involving CPU/GPU architectures, heterogeneous computing, AI/ML compilation, HPC, parallel programming, or performance optimization.
- Experience with AI frameworks such as PyTorch, inference engines such as vLLM and SGLang, and machine-learning models such as DeepSeek, Hugging Face, and Kimi.