CPU Compiler Intern

About the role

NVIDIA’s Accelerated Computing team is looking for a motivated intern to work on CPU compiler development and high‑performance numerical libraries. The role is based in Pune (with an additional opening in Bengaluru) and requires a full‑time commitment of at least 12 weeks.

Responsibilities

• Develop and enhance MLIR‑based LLVM compiler components for CPU architectures.
• Design and implement efficient algorithms for accelerated numerical computing.
• Collaborate with cross‑functional compiler, software, and hardware teams to deliver high‑quality code.
• Engage with the open‑source MLIR and LLVM communities to contribute improvements.
• Validate and benchmark compiler performance on NVIDIA’s CPU platforms.

Eligibility

• Currently enrolled in a Master’s or PhD program in Computer Engineering, Electrical Engineering, Computer Science, Mathematics or a related discipline.
• Expected graduation between April 2027 and September 2027.
• 0–1 years of prior internship or full‑time experience; more than one year disqualifies the candidate.
• Authorized to work in India without NVIDIA sponsorship.
• Ability to commit to a minimum 12‑week full‑time internship.

Skills

• Strong understanding of compiler design, runtimes and programming language concepts.
• Proficiency in C++ and Python.
• Familiarity with LLVM, MLIR, and GPU‑related toolchains such as CUDA, cuDNN, cuBLAS.
• Knowledge of CPU architecture, parallel programming, algorithms and numerical methods.

Benefits

• Industry‑leading stipend (exact figure undisclosed).
• Hands‑on experience with cutting‑edge compiler and AI hardware technologies.
• Mentorship from senior engineers in NVIDIA’s Accelerated Computing division.
• Opportunity to contribute to open‑source compiler projects.
• Access to NVIDIA’s internal resources and learning platforms.