Principal Engineer, Compilers and Formal Methods

at Nvidia

📍 Seattle, United States

USD 248,000-391,000 per year

SENIOR

✅ On-site

Tech Stack
Tag name is followed by "@" symbol and proficiency level value. About proficiency levels:

1-2 — basic awareness. Minimal hands-on experience, and a rudimentary understanding of the technology's purpose;

3-6 — daily use. Comfortable and regular usage, capable of handling common tasks and challenges related to the technology;

7-9 — you are an expert, you can teach others, you know all the pitfalls and tricks;

10 — exceptional knowledge, comprehensive understanding, and adeptness in all aspects of the technology, including advanced problem-solving. Think twice before claiming or demanding such level.

AI @ 6 CUDA Debugging GPU GenAI Generative AI LLM Mathematics @ 4

Details

NVIDIA is seeking an exceptional computer scientist to lead research at the intersection of compilers, programming languages, formal methods, and automated reasoning. The role involves developing next-generation compiler technologies that combine advanced program analysis and optimization with rigorous methods for establishing program correctness. It spans fundamental research and practical compiler engineering, with opportunities to influence NVIDIA's compiler stack for GPUs, AI accelerators, domain-specific systems, and emerging computing architectures.

Responsibilities

Lead research in compiler technology, programming languages, formal methods, and automated reasoning.
Develop next-generation compiler techniques for translating high-level programs into efficient code for GPUs and heterogeneous computing platforms.
Develop formal foundations for compiler transformations and establish that optimizations preserve specified program semantics.
Research verified and formally grounded compiler optimizations, including techniques for automatically proving the correctness of transformations.
Develop scalable program analyses that enable compilers to reason about increasingly complex programs and transformations.
Investigate techniques for automatically generating proofs, invariants, specifications, and correctness certificates for compiler transformations.
Investigate how large language models and generative AI can assist compiler construction, optimization, program transformation, verification, and debugging.
Collaborate with NVIDIA compiler, CUDA, GPU architecture, AI, systems software, and research teams to transition research prototypes into production compiler technology.
Mentor researchers and engineers working on compiler technology, programming languages, formal verification, and automated reasoning.

Requirements

PhD in Computer Science, Computer Engineering, Mathematics, or a related field, or equivalent research experience.
15+ years of academic and/or industry expertise in compiler construction, programming languages, formal methods, or closely related areas.
Strong understanding of compiler fundamentals, including intermediate representations, control-flow and data-flow analysis, static analysis, and program transformation.

Preferred Qualifications

Research experience in verified compilation or formally verified compiler transformations.
Experience with translation validation, certified compilation, proof-producing compilation, and semantics-preserving optimization.
Background with SMT/SAT solvers and automated reasoning systems.
Experience developing sophisticated compiler infrastructure or intermediate representations.

Benefits

NVIDIA offers competitive salaries, equity, and a comprehensive benefits package. NVIDIA is committed to fostering an inclusive work environment and is an equal opportunity employer.