AMD seeks a Lead AI Compiler Engineer in Hyderabad to enhance on-device AI inference acceleration. The role involves designing AI compiler stacks, optimizing kernels, and transforming AI models for AMD NPUs. Strong C/C++, Python, and AI compiler expertise required. Join AMD’s cutting-edge AI team today.
Location: Hyderabad, India
Job ID: 64069
Category: Engineering
Apply: Click Here
Operator Fusion: Analyze and develop optimizations that reduce memory traffic while doing AI model operation.
Problem Partitioning & Dataflow Orchestration: Assign AI operations to NPUs and develop efficient dataflow through multi-level memory hierarchies.
Kernel Design & Development: Write high-performance C++/intrinsic kernels leveraging SIMD and VLIW architectures.
Testing & Validation: Develop 'model' representation on CPU, as well as unit tests and integration tests to verify correctness and confidence in results.
Performance Profiling & Tuning: Analyze each layer in model, identify performance bottlenecks, and optimize accuracy, memory and computation trade-offs.
Documentation & Collaboration: Document design specifications, follow coding best practices, and cooperate with AI research, architecture, and software teams.
Proficient in C/C++ and Python.
Familiar with SIMD, VLIW, vectorized programming and fixed/floating point arithmetic.
Inspired to learn about state-of-the-art AI operators and low-level hardware details (caching and DMA programming).
Experience in AI compiler design is a plus.
Strong problem solving, debugging and performance improvement capabilities.
Bachelor’s, Master’s, or PhD in Computer Science, Electrical Engineering, or related field.
AMD (Advanced Micro Devices) is a global leader in high-performance computing and graphics solutions, fueling innovation in AI, data center, gaming, PCs, and embedded systems. AMD is committed to bringing people together through collaboration and inclusion, while creating a better world for all by shaping the future of adaptive computing and communities around the world.