GPU Compiler Backend Engineer (Pune)

Designation - GPU Compiler Backend Engineer Experience Level: 8+ Years Role Type: Senior Individual Contributor / Technical Lead Location: Pune / Bangalore / Hyderabad We are looking for a highly skilled GPU Compiler Backend Engineer who can design and optimize LLVM- or GCC-based backend compiler components for massively parallel GPU architectures. This role is focused on building high-performance code generation paths for compute, graphics, and HPC workloads, with direct impact on GPU and SoC products. The ideal candidate has deep compiler experience, strong C/C++ fundamentals, and a solid understanding of GPU or CPU architecture. Key Responsibilities Design and implement backend compiler features for GPU targets using LLVM or GCC infrastructure. Develop new compiler passes and optimizations for performance, compile-time efficiency, and code quality. Work on instruction selection, instruction scheduling, register allocation, lowering, and machine-level optimizations. Improve support for LLVM IR, Machine IR (MIR), TableGen, and target-specific code generation. Analyze and resolve performance bottlenecks, codegen bugs, miscompilations, and runtime issues. Collaborate with architecture, hardware, driver, and application teams to align compiler behavior with platform goals. Benchmark and tune workloads across Deep Learning, Graphics, HPC, and compute pipelines. Contribute to design reviews, technical roadmap discussions, and cross-team problem solving Required Technical Skills Deep expertise in C and C++. Strong hands-on experience with LLVM or GCC compiler internals. Solid background in compiler design and code generation. Knowledge of GPU or CPU architecture, pipeline behavior, and low-level performance trade-offs. Understanding of instruction scheduling, register allocation, software pipelining, and backend lowering concepts. Experience with performance analysis tools, profiling, and debugging complex software issues. Strong Linux development skills, Git, build systems, and automated testing. Preferred Skills GPU compiler backend development experience. Exposure to CUDA, OpenGL, Vulkan, DirectX, OpenCL, or similar workloads. Familiarity with GlobalISel, TableGen, LLVM IR, and MIR. Experience tuning compilers for parallel or vector architectures. Exposure to AI/ML compilation, graph optimizations, or inference stack integration. Open-source compiler contribution experience. Education B.E. / B.Tech / M.E. / M.Tech / Ph.D. in Computer Science, Electronics, Electrical Engineering, or related field. Equivalent hands-on compiler engineering experience is acceptable for exceptionally strong candidates. Project Experience Compiler backend or low-level toolchain development on production systems. Work on code generation, performance tuning, or architecture-specific optimization. Experience debugging issues across compiler, runtime, and hardware boundaries. Strong evidence of building maintainable systems software at scale. Soft Skills Strong analytical depth and persistence in debugging difficult problems. Excellent communication and technical documentation skills. Ability to influence architecture and engineering decisions. Self-motivation and ownership of complex technical work. Valuable to Have Direct experience with GPU architecture teams. Familiarity with deep learning compilers or graph lowering stacks. Experience with compiler validation, benchmarks, and regression systems. Contributions to open-source compilers or code generation infrastructure. Expected Deliverables Backend compiler features and optimizations. Robust, maintainable code generation improvements. Benchmark gains in target workloads. Regression coverage and technical documentation. Cross-functional technical recommendations for future GPU architectures. .