GPU (Graphics) - AI Performance Engineer

This role is for one of the Weekday's clients

Salary range: Rs 3000000 - Rs 10000000 (ie INR 30-100 LPA)

Experience: 6+ yrs

Location: Bengaluru

Job Type: full-time

We are seeking an experienced Graphics/AI Performance Engineer to drive power and performance optimization for next-generation GPU architectures and AI workloads. This role is ideal for professionals with deep expertise in GPU architecture, performance analysis, and power modelling who are passionate about building high-performance, energy-efficient computing platforms.

As a Graphics/AI Performance Engineer, you will work closely with architecture, hardware, software, compiler, driver, and machine learning teams to analyze workload behavior, identify system bottlenecks, and optimize GPU performance across graphics and AI applications. You will play a key role in developing power-performance models, evaluating architectural trade-offs, and influencing future GPU design decisions through data-driven insights. This position offers the opportunity to contribute to cutting-edge technologies at the intersection of graphics, AI, and semiconductor engineering.

• Develop and maintain GPU power and performance models for graphics, compute, and AI workloads.
• Analyze GPU architecture and microarchitecture to identify opportunities for performance improvements and power optimization.
• Evaluate AI and graphics workloads to understand execution behavior and correlate performance metrics with power consumption.
• Perform workload characterization, benchmarking, and profiling to identify system bottlenecks and optimization opportunities.
• Conduct power-performance trade-off studies and provide data-driven recommendations for architectural enhancements.
• Collaborate with architecture, hardware, software, compiler, driver, and machine learning teams to optimize end-to-end system performance.
• Identify opportunities for hardware-software co-optimization to improve efficiency across graphics and AI applications.
• Utilize GPU computing technologies and performance analysis tools to evaluate application behavior and optimize execution.
• Support future GPU architecture planning by providing insights based on performance modelling and workload analysis.
• Develop automation scripts and analysis tools using C/C++ and Python to improve productivity and performance evaluation workflows.
• Document technical findings, optimization strategies, and architectural recommendations for engineering teams.

• Bachelor's degree with 6+ years, Master's degree with 5+ years, or PhD with 4+ years of relevant experience in performance engineering or semiconductor design.
• Strong expertise in GPU Architecture, GPU microarchitecture, and GPU Computing technologies.
• Hands-on experience with power analysis methodologies and tools such as PrimeTime PX, Power Artist, or equivalent solutions.
• Strong understanding of performance analysis, workload characterization, benchmarking, and system optimization techniques.
• Experience working with AI/ML frameworks such as TensorFlow or PyTorch for workload analysis and optimization.
• Proficiency in C/C++ and Python for performance modelling, scripting, and automation.
• Familiarity with GPU computing APIs and frameworks such as CUDA, OpenCL, Vulkan, OpenGL, or DirectX.
• Knowledge of low-power ASIC design techniques and hardware-software co-optimization principles.
• Understanding of GPU drivers, compiler technologies, runtime software, and performance profiling methodologies is an advantage.
• Exposure to Verilog/SystemVerilog and semiconductor design flows is desirable.
• Strong analytical, problem-solving, and communication skills with the ability to collaborate effectively across multidisciplinary engineering teams.