GPU Compiler Development Engineer

NVIDIA; Maharashtra; On-site; INR 1,200,000 - 1,900,000; Full time

Job summary

NVIDIA is seeking an experienced compiler engineer for its GPU Assembler team in India. You will develop and optimize the Assembler and Disassembler for GPU compute, collaborating with architecture, debugger, and profiler teams to deliver high-performance code.

The role requires 2+ years in compiler or low-level tool development, strong C/C++ skills, and a solid grasp of ISA and architecture. You will work independently and drive key features from concept to release.

Qualifications

• BS or MS degree in Computer Science/Engineering or equivalent.
• At least 2+ years in software development related to assemblers, linkers, loaders, and debuggers.
• Strong C/C++ programming and understanding of Assembly language/ISA.

Responsibilities

• Develop, maintain, and improve NVIDIA's Assembler and Disassembler for GPU compute.
• Collaborate with architecture and debugger teams to deliver vital features.
• Ensure compatibility and high-quality code generation across related components.
• Engage with customers to gather feedback and drive feature ideas.

Skills

C/C++ programming

Analytical skills

Assembly language

Computer Architecture

Independent work

Education

BS or MS in CS/CE

Tools

Disassembler tools

Binary tools

GDB

BOLT

Job description

We are looking for an experienced compiler engineer for an exciting role in the GPU Assembler team. Join the GPU Assembler team and help build tools that enable creation of the most performant applications and libraries for HPC, DL and Autonomous Driving domain. Work with key HPC developers around the world to help them achieve the best performance & efficiency on current and next generation NVIDIA GPUs by developing best in industry Assembler and Code Analysis Tools.

We will be part of a team that develops a core component of our platform for accelerating general purpose computation on the GPU. Disassembler (NVDISASM) developed by the team is used by all the CUDA users worldwide to analyse their binaries with various code analysis features built into the disassembler. Assembler is used for achieving the best performance for key DL applications.

What You Will Be Doing

• Develop, maintain, and consistently improve NVIDIA’s Assembler and Disassembler for GPU compute, ensuring they meet the highest standards of performance and reliability.
• Collaborate with GPU architecture and debugger/profiler development teams to understand their requirements and deliver vital features and improvements.
• Work closely with teams developing related components to ensure compatibility, reliability, and high-quality code generation.
• Engage with customers and partners to gather feedback, driving innovative ideas and features that improve our product's capabilities.

What We Need To See

• BS or MS degree in Computer Science, Computer Engineering, or equivalent experience.
• At least 2+ years of software development experience related to assemblers, linkers, loaders, binary tools, and debuggers.
• Strong analytical skills and proficiency in C/C++ programming.
• Experience in any area of compiler development, including feature support, code generation, and compiler infrastructure.
• Solid understanding of Assembly Language / Processor ISA, Computer Architecture, and system software.
• Ability to work independently, define task goals and scope, and lead development efforts successfully.

Ways To Stand Out From The Crowd

• Deep understanding of debugger/profiler tools, bintools, linker internals, and experience with binary analysis/instrumentation tools like BOLT.
• Experience using AI tools in everyday work such as Cursor, Windsurf, etc.
• Knowledge of GPU development and compute APIs like CUDA and OpenCL.

Join us at NVIDIA where your outstanding skills and innovative outlook will be celebrated and nurtured. Be part of a team that is not just crafting the future but shaping it. This is your opportunity to take your career to the next level and be part of something truly groundbreaking!