Principal Engineer, Applied Research - Accelerator Programming Model and Compiler

Job Title

Principal Engineer, Applied Research - Accelerator Programming Model and Compiler

Role Summary

Design and lead the programming model, SDK, and LLVM-based compiler backend for NVIDIA’s Programmable Vision Accelerator (PVA). The role sits between applied research and production systems software and focuses on enabling developer- and AI agent–driven workflows for power-efficient VLIW/SIMD accelerators used across NVIDIA DRIVE, Jetson, and IGX platforms.

Experience Level

Senior — requires extensive experience (15+ years) in high-performance systems, compilers, or accelerator software.

Responsibilities

Deliver architecture, tooling, and integrations that make PVA workloads easier to express, generate, and optimize for both humans and AI agents.

• Define programming model and runtime APIs for PVA targeting both developers and AI coding agents.
• Design and evolve the PVA SDK, compiler interfaces, and diagnostic tooling.
• Develop and improve an LLVM-based VPU compiler backend for a VLIW/SIMD architecture.
• Abstract hardware details into declarative interfaces and DSL/IR targets for AI-driven code generation.
• Integrate PVA workloads into CUDA-based heterogeneous pipelines (execution, memory movement, synchronization).
• Define benchmarks and evaluate agent-generated code for correctness and performance; close compiler-feedback loops.
• Collaborate with internal and external customers to drive adoption and shape real-world runtime APIs.

Requirements

Must-have technical skills and experience. Nice-to-have items listed separately.

• 15+ years building high-performance, low-level systems software, accelerator software, embedded software, or compiler/toolchain infrastructure.
• Experience building higher-level programming abstractions, DSLs, or compiler IRs that preserve performance while hiding hardware complexity.
• Experience with compiler, debugger, linker, or toolchain development, particularly using LLVM.
• Experience developing with AI coding agents and integrating compilers/tools with agentic development harnesses.
• Experience programming SIMD/VLIW processors and strong C++ development skills, including low-level debugging and profiling.
• Experience with Linux or QNX development environments.
• Strong communication and collaboration skills.

Nice-to-have:

• Experience with CUDA and integrating accelerators into CUDA-based heterogeneous compute pipelines.
• Familiarity with MLIR, Halide, TVM, Triton, graph compilers, or image-processing DSLs.
• Familiarity with ISO 26262, IEC 61508, or equivalent safety/quality standards.
• Experience building agent harnesses and orchestrating long-running, stateful multi-agent workflows.

Education Requirements

BS, MS, or PhD in Computer Science, Computer Engineering, Electrical Engineering, or a related field, or equivalent practical experience.

About the Company

Company: NVIDIA

Headquarters: Santa Clara, California, USA

NVIDIA is a global leader in accelerated computing, renowned for its innovative solutions in AI and digital twins that transform diverse industries. The company specializes in networking technologies, providing end-to-end InfiniBand and Ethernet solutions for servers and storage that optimize performance and scalability. NVIDIA serves sectors such as high-performance computing, enterprise data centers, and cloud computing, constantly reinventing its products and services to stay ahead in the market.