GPU System Architect

Position Overview

We are seeking an experienced GPU System Architect to define, design, and prototype the architecture of the company's next-generation GPU products. The role will lead key microarchitecture decisions across GPU compute, cache and memory subsystems, and graphics/AI compute pipelines, with clear ownership of performance, power, and area (PPA) targets. The architect will drive the product from architectural definition and performance modeling through FPGA prototyping, hardware-software co-verification, and implementation support toward tape-out. This position is based in Singapore and requires close day-to-day technical collaboration with Chinese-speaking engineering teams.

Core Responsibilities

1. Lead the exploration and definition of next-generation GPU system and core architectures, including compute architecture, cache and memory subsystems, and graphics/AI compute pipelines.

2. Define key microarchitecture solutions and produce clear architecture specifications covering module functionality, interfaces, design assumptions, performance requirements, and PPA targets.

3. Build and maintain architecture-level performance simulators and C/C++- or Python-based analytical models to evaluate representative workloads including AI inference/training, graphics rendering, and scientific or general-purpose computing.

4. Perform quantitative architecture trade-off studies, identify performance and power bottlenecks, and recommend design choices that achieve the best overall product balance.

5. Own the FPGA prototype validation workflow: define prototype strategies, map key architectural modules to FPGA platforms, and resolve issues involving timing closure, resource utilization, interface adaptation, and system bring-up.

6. Partner closely with low-level software teams on drivers, firmware, operators/kernels, and prototype-platform debugging to enable early hardware-software co-verification and rapid architecture iteration based on measured results.

7. Work with RTL design, functional verification, and physical implementation teams to communicate architectural intent, resolve implementation issues, and ensure the final design remains aligned with architectural and PPA goals.

8. Lead technical reviews and cross-team architectural decisions, drive issue closure across the product lifecycle, and support the transition from prototype validation to tape-out readiness.

9. Track global GPU and accelerator technology trends, evaluate relevant architectural innovations, and develop forward-looking technical options for future product generations.

10. Collaborate effectively across Singapore and Chinese-speaking engineering teams, ensuring complex architecture decisions, assumptions, and technical risks are communicated clearly and consistently.

Qualifications

· Bachelor's degree or above in Microelectronics, Computer Science, Electronic Engineering, Computer Engineering, or a related discipline.

· Approximately 10 years of relevant industry experience in architecture design for high-performance GPUs, general-purpose computing chips, AI accelerators, or closely related high-performance processor architectures.

· Strong foundation in computer architecture and a deep understanding of GPU operating principles; familiarity with the architectural characteristics of at least one mainstream commercial GPU architecture.

· Broad system-level understanding across GPU compute pipelines, cache/memory hierarchy, graphics or AI compute pipelines, and the interactions between major architectural subsystems.

· Strong proficiency in C/C++ and Python, with proven ability to build architecture-level models, performance simulators, and quantitative evaluation frameworks independently.

· Hands-on FPGA prototype validation experience and familiarity with mainstream AMD/Xilinx or Intel FPGA development flows; capable of independently troubleshooting complex prototype-mapping and bring-up issues.

· Demonstrated experience working with low-level software, RTL design, verification, and physical implementation teams on hardware-software co-design and architecture implementation.

· Professional proficiency in Mandarin Chinese is required for technical discussions and day-to-day collaboration with Chinese-speaking engineering teams. Good working proficiency in English is also required for the Singapore-based and cross-functional working environment.

· Strong technical leadership, structured problem-solving, and communication skills, with the ability to explain complex architectural concepts clearly and drive decisions across multiple teams.

· Self-motivated and execution-oriented, with the ability to work effectively in a fast-iteration environment and lead complex technical projects under practical resource and schedule constraints.

Preferred Qualifications

· Prior experience in AI accelerator or heterogeneous compute architecture design.

· Experience leading a GPU or high-performance processor architecture from definition through prototype bring-up and tape-out support.

· Experience mentoring engineers or leading a small architecture/technical team.

· Familiarity with performance optimization across hardware and low-level software stacks is advantageous.