Senior Deep Learning Compiler Engineer - PyTorch

Overview

Join a pioneering team at the cutting edge of AI compiler innovation to help revolutionize accelerated computing. As part of NVIDIA, you will develop Thunder, a state-of-the-art source-to-source compiler designed to maximize the performance of PyTorch models on NVIDIA GPUs. This role presents a rare chance to influence the PyTorch ecosystem by integrating advanced modern compiler technologies including PyTorch 2.0's TorchDynamo and TorchInductor, generating impactful open-source tools benefiting the global AI research community.

Key Responsibilities

In this dynamic and technically demanding position, you will spearhead the architecture, implementation, optimization, and upkeep of core compiler technologies that drive large-scale deep learning tasks. Collaborating closely with the original PyTorch engineers at NVIDIA, you will spearhead innovative features, analyze performance across thousands of GPUs, and identify optimization avenues shaping Thunder’s future. You will also partner with cross-functional teams specializing in nvFuser, TVM, XLA, and CUDA to transform cutting-edge research into practical, high-impact solutions for the broader open-source ecosystem.

Required Qualifications

• A Bachelor’s, Master’s, or Ph.D. in Computer Science or a related domain, or equivalent professional experience.
• Over 8 years of relevant industry experience.
• Proficiency in Python and a track record of developing complex, well-tested software systems.
• Hands-on experience with deep learning frameworks such as PyTorch or JAX, with a clear understanding of model construction and performance bottlenecks.
• Strong knowledge of compiler theory including abstract syntax trees, intermediate representations like SSA form, program analysis, and code generation.
• Excellent communication and team collaboration abilities, especially in distributed and open-source development contexts.

Preferred Distinctions

• Contributions to deep learning compiler projects (e.g., TVM, MLIR, IREE) or deep learning frameworks.
• In-depth understanding of PyTorch internals, especially its compiler stack (TorchDynamo, TorchInductor).
• Experience working with JAX-style functional transformations within compiler frameworks.
• Knowledge of parallel programming, distributed systems, and CUDA-based high-performance computing.
• Active involvement in open-source communities through code, design discussions, or mentorship.

Additional Information

NVIDIA is a leader in Artificial Intelligence, High-Performance Computing, and Visualization, offering competitive compensation, extensive benefits, and a workplace that encourages diversity, inclusion, and flexibility. This ensures a nurturing environment for all employees. Salary is based on location, candidate experience, and comparable role compensation. Specifically for Poland, the base salary range is between 292,500 PLN and 507,000 PLN annually.