Chief GPU Architect

Job Description A global technology leader is seeking a Chief GPU Architect / Research Lab Director to guide the architectural vision for next-generation mobile and heterogeneous computing platforms.This role will lead the exploration and development of advanced compute architectures where graphics processing and AI acceleration converge within a unified subsystem. The position requires deep expertise in GPU design, heterogeneous computing, and large-scale silicon product development, along with the ability to translate emerging workloads into practical architectural solutions.The selected leader will define architectural strategies that enable future mobile platforms to efficiently support complex graphics and AI workloads, while maximizing performance efficiency, scalability, and silicon utilization.ResponsibilitiesInvestigate future graphics and AI workloads to inform long-term architectural direction.Develop compute architectures that integrate graphics processing capabilities with machine learning acceleration.Evaluate approaches for combining dedicated hardware accelerators with shared programmable compute resources.Design execution units capable of supporting mixed workloads, including vector, scalar, and tensor operations.Create scheduling models and dispatch strategies capable of efficiently handling diverse compute workloads.Define mechanisms that allow compute resources to be shared dynamically while maintaining performance isolation where required.Establish the technical roadmap for upcoming mobile GPU generations.Lead architectural innovation across shader cores, compute pipelines, and memory subsystems.Balance key design goals including performance, power efficiency, and silicon area optimization.Provide architectural guidance from early research concepts through implementation and post-silicon optimization.Define scalable memory architectures tailored for heterogeneous compute environments.Design shared caching approaches and efficient bandwidth management techniques.Improve on-chip communication infrastructure to support diverse compute traffic patterns.Reduce latency and power consumption associated with data movement across compute components.Work closely with CPU architects, AI software engineers, runtime developers, and SoC design teams.Contribute to overall system architecture decisions involving power budgets, thermal limits, and physical layout.Ensure hardware architecture aligns with evolving graphics APIs, machine learning frameworks, and runtime environments.Support modelling, benchmarking, and performance analysis across development stages.Required Experience15+ years of experience in GPU architecture, heterogeneous compute design, or AI accelerator development.Strong understanding of GPU execution models such as SIMD/SIMT architectures.Expertise in compute scheduling, pipeline design, and memory hierarchy optimization.Experience with tensor or matrix computation used in modern machine learning workloads.Demonstrated track record delivering complex semiconductor products to production.Proficiency in performance modelling, workload analysis, and power optimization techniques.Leadership experience guiding multidisciplinary engineering teams and influencing technical direction.Experience designing architectures that combine graphics and machine learning compute capabilities.Knowledge of mobile or energy-constrained computing environments.Familiarity with modern AI software ecosystems and graphics programming interfaces.Experience managing architecture groups, advanced research teams, or innovation labs.If you are motivated by the opportunity to shape next-generation compute architectures and believe your experience aligns with this role, we would be pleased to hear from you.Qualified and interested candidates are invited to submit their CV for confidential consideration to [email protected] applying to this role you understand that we may collect your personal data and store and process it on our systems. For more information please see our Privacy Notice (https://eu-recruit.com/about-us/privacy-notice/)

Created: 2026-03-20