– Senior SoC Architect (GPU Architecture)

Experience: Senior / Lead Level (Typically 10+ Years)

Domain: Semiconductor / VLSI / SoC Architecture

Specialization: GPU Architecture and Microarchitecture

Job Location: Bengaluru

Employment Type: Full time

- Job Overview

We are looking for an experienced Senior SoC Architect with deep expertise in GPU architecture, microarchitecture and complex SoC design. The ideal candidate should have a strong background in designing and architecting high-performance GPU-based SoCs, with hands-on involvement in defining GPU architecture, microarchitecture specifications, system-level integration and performance optimization.

The candidate will be responsible for driving architectural decisions for next-generation GPU solutions, including GPU compute engines, shader processing units, graphics pipelines, memory hierarchy, cache architecture, interconnect fabrics and system-level interfaces.

The role requires a strong understanding of GPU architecture, parallel processing, high-performance computing, memory subsystems, power-performance-area (PPA) optimization and advanced semiconductor technologies.

The candidate will work closely with RTL design, verification, physical design, software and platform teams to translate architectural requirements into implementable silicon solutions.

- Key Responsibilities

A. GPU Architecture and Microarchitecture
- Define and develop GPU architecture for high-performance computing, graphics processing and parallel computing applications.
- Drive architectural specifications for GPU IPs and GPU-based SoC platforms.
- Define GPU microarchitecture covering execution units, shader cores, compute units, processing clusters and graphics processing pipelines.
- Architect GPU execution pipelines, including vertex processing, geometry processing, rasterization, pixel shading and compute workloads.
- Define architecture for SIMD/SIMT processing, vector execution, parallel execution and thread scheduling.
- Develop architectural specifications for shader cores, ALUs, floating-point units, integer execution units and special function units.
- Evaluate different GPU architectural approaches to improve performance, scalability, throughput and power efficiency.
- Define GPU architectural requirements based on target workloads, application requirements and performance benchmarks.
- Analyze and optimize GPU architecture for graphics rendering, AI/ML acceleration, HPC and general-purpose GPU computing.

B. GPU Compute Architecture
- Define GPU compute architecture, including compute units, processing elements and execution pipelines.
- Architect parallel processing capabilities for high-throughput computing workloads.
- Define thread execution models, warp/wavefront scheduling and execution management.
- Develop architecture for instruction issue, dispatch, scheduling and execution.
- Define execution models for SIMD, SIMT and vector processing.
- Evaluate compute throughput, instruction-level parallelism and execution efficiency.
- Optimize GPU resource utilization and workload distribution across compute units.
- Work on architectural improvements for compute-intensive workloads, including AI/ML and scientific computing.

C. Memory Subsystem and Cache Architecture
- Define and optimize GPU memory hierarchy, including L1, L2 and system-level caches.
- Architect memory controllers, memory interfaces and memory access mechanisms.
- Define cache coherency,



memory consistency and data-sharing mechanisms.
- Work on memory bandwidth optimization, latency reduction and efficient data movement.
- Define memory subsystem architecture for high-performance GPU workloads.
- Evaluate and architect memory interfaces such as:
- HBM / HBM2 / HBM3
- GDDR
- LPDDR
- DDR
- Optimize memory bandwidth, cache utilization and memory access efficiency.
- Define architecture for shared memory, local memory, global memory and register files.
- Work on memory partitioning, banking, arbitration and memory scheduling.

D. SoC Architecture and Integration
- Define overall SoC architecture integrating GPU IPs, CPU subsystems, memory controllers, interconnects and peripheral interfaces.
- Develop architectural specifications and high-level block diagrams for complex SoC designs.
- Define communication architecture between GPU, CPU, memory and other SoC subsystems.
- Architect high-performance interconnect fabrics and NoC (Network-on-Chip) solutions.
- Define data movement architecture and communication protocols between various processing elements.
- Drive IP integration strategy and interface specifications.
- Evaluate system-level performance bottlenecks and recommend architectural improvements.
- Work with cross-functional teams to ensure architectural feasibility and successful silicon implementation.

E. High-Speed Interfaces and Interconnect
- Define high-bandwidth communication interfaces for GPU and SoC integration.
- Architect high-speed interfaces such as PCIe, CXL, AXI, AMBA and other industry-standard protocols.
- Work on high-speed data transfer between GPU, CPU, memory and external devices.
- Define interconnect topology, bandwidth requirements, arbitration and routing strategies.
- Optimize communication latency and bandwidth utilization across the SoC.
- Evaluate scalability and performance of interconnect architecture.

F. Power, Performance and Area Optimization
- Drive architectural optimization for power, performance and area (PPA).
- Perform architectural trade-off analysis to balance performance, power consumption and silicon area.
- Define power management strategies for GPU compute units and SoC subsystems.
- Work on dynamic voltage and frequency scaling (DVFS), clock gating and power gating methodologies.
- Identify performance bottlenecks through architectural analysis and workload profiling.
- Perform system-level performance modeling and architecture evaluation.
- Collaborate with physical design teams to understand implementation constraints and optimize architecture accordingly.

G. Architecture Definition and Technical Leadership
- Own and drive architectural specifications, microarchitecture documents and technical design requirements.
- Lead architecture discussions and make critical design decisions.
- Evaluate new GPU technologies, architectural trends and emerging semiconductor solutions.
- Collaborate with RTL, verification, physical design, software and system engineering teams.
- Review RTL and microarchitecture implementations against architectural specifications.
- Drive design reviews,



technical discussions and architectural signoff.
- Mentor and guide engineers working on GPU and SoC architecture.
- Support pre-silicon validation, post-silicon debugging and silicon bring-up activities.

3. Required Technical Skills GPU Architecture – Mandatory
- Strong experience in GPU architecture and microarchitecture.
- Deep understanding of GPU processing pipelines and parallel computing architecture.
- Experience with shader cores, compute units, execution engines and GPU processing clusters.
- Strong knowledge of SIMD/SIMT execution models.
- Understanding of GPU scheduling, dispatch, execution and resource management.
- Experience in GPU performance analysis and architectural optimization.

SoC Architecture – Mandatory
- Strong experience in SoC architecture and system-level design.
- Experience in defining complex SoC architectures and subsystem integration.
- Understanding of CPU-GPU integration and heterogeneous computing architecture.
- Knowledge of AXI, AMBA and NoC-based interconnect architectures.
- Experience with memory subsystem and cache architecture.
- Ability to define architecture specifications and microarchitecture requirements.

Memory and Interconnect
- Strong understanding of cache hierarchy, memory controllers and memory bandwidth optimization.
- Experience with DDR, LPDDR, GDDR or HBM memory architectures.
- Understanding of cache coherency and memory consistency protocols.
- Knowledge of PCIe, CXL, AXI and high-speed interconnect technologies.

Performance and Power Optimization
- Experience in architectural performance modeling and analysis.
- Understanding of power-performance-area trade-offs.
- Knowledge of DVFS, clock gating and power management techniques.
- Experience in identifying architectural bottlenecks and optimizing system performance.

4. Preferred Technical Skills
- Experience in GPU IP development or GPU-based SoC development.
- Knowledge of graphics APIs and GPU programming models such as Vulkan, OpenGL, DirectX, CUDA or OpenCL.
- Understanding of graphics rendering pipelines and shader programming concepts.
- Experience with AI/ML acceleration and heterogeneous computing.
- Knowledge of RISC-V or ARM-based SoC architecture.
- Exposure to advanced semiconductor technology nodes such as 7nm, 5nm, 3nm or below.
- Experience in RTL design methodologies and hardware description languages such as Verilog/SystemVerilog.
- Understanding of formal verification, simulation and emulation methodologies.
- Exposure to silicon bring-up, post-silicon validation and performance characterization.

5. Educational Qualification
- B.E./B.Tech/M.Tech in Electronics, Electrical Engineering, Computer Engineering, VLSI or related disciplines.
- M.S./Ph.D. in Computer Architecture, VLSI or related areas will be an advantage.

6. Expected Candidate Background Candidates should ideally have experience in one or more of the following areas:
- GPU Architecture and Microarchitecture.
- GPU IP Development.
- GPU Compute Architecture.
- Graphics Processing Architecture.
- High-Performance SoC Architecture.
- CPU-GPU Heterogeneous Computing.
- GPU Memory Subsystem and Cache Architecture.
- NoC and SoC Interconnect Architecture.
- AI/ML Accelerator Architecture.

Key Expectation: The candidate should have substantial hands-on involvement in GPU architecture and should be capable of driving architecture definition, evaluating architectural trade-offs and leading complex GPU/SoC design decisions.

📌 Senior SOC Architect (Bengaluru)
🏢 leadIC Design
📍 Bengaluru