Senior AI Architect, Foundation Models and SoC Co-Design – Autonomous Vehicles

NVIDIA•Us, CA

36d

About The Position

NVIDIA is at the forefront of accelerated computing, AI, and autonomous machines. From generative AI to robotics and self-driving vehicles, our technologies are transforming some of the world’s largest industries. NVIDIA DRIVE™ is redefining autonomous mobility through state-of-the-art AI, high-performance compute, and scalable software-defined architectures. We are looking for a Senior AI Architect to help define the next generation of AI model paradigms for autonomous vehicles and shape how those models co-evolve with NVIDIA’s future embedded SoC architectures. This is a highly strategic role operating at the intersection of frontier AI research, hardware architecture, systems optimization, and autonomous driving. You will work with world-class AI researchers, silicon architects, and AV platform teams to identify the AI workloads that will define the next decade — and ensure NVIDIA platforms are architected to lead them.

Requirements

MS, PhD, or equivalent experience in Computer Science, Electrical Engineering, Machine Learning, Robotics, or related field.
12+ years of experience in AI/ML systems, deep learning architecture, or hardware/software co-design.
Deep expertise in modern AI architectures and large-scale model systems
Experience mapping AI workloads onto heterogeneous compute architectures including GPUs, CPUs, NPUs/DLAs, DSPs, and memory subsystems.
Solid understanding of distributed training systems, scaling laws, and inference optimization techniques.
Experience with model optimization methods such as quantization, sparsity, pruning, distillation, and memory-efficient inference.
Understanding of performance profiling, systems bottleneck analysis, and workload characterization.

Nice To Haves

Experience with autonomous vehicle or robotics stacks including perception, planning, prediction, or control.
Deep familiarity with NVIDIA platforms such as DRIVE™, Jetson™, CUDA®, TensorRT™, Triton, or TensorRT-LLM.
Experience influencing silicon architecture or collaborating directly with hardware design teams.
Expertise in sophisticated AI efficiency techniques (e.g. FP8/FP4 inference, Mixture-of-Experts routing, Streaming attention and KV-cache optimization)
Strong understanding of multimodal fusion across camera, lidar, radar, HD maps, and language inputs.

Responsibilities

Research and forecast emerging AI model architectures that are expected to shape the future autonomous vehicle stack, including Vision-Language-Action (VLA) models, Multimodal foundation models and more.
Drive hardware-software co-design across next-generation AI workloads and NVIDIA embedded SoCs, including GPU, CPU, DLA, memory hierarchy, interconnects, and accelerator subsystems.
Analyze compute, memory, bandwidth, and latency characteristics of sophisticated AI architectures such as transformers, diffusion models, or MoE systems
Develop architectural insights and influence future NVIDIA silicon, IP, and system-level design decisions through deep workload characterization and performance analysis.
Prototype and evaluate emerging model paradigms on NVIDIA DRIVE and embedded AI platforms to validate scalability, efficiency, and deployment feasibility.
Partner closely with AI research, autonomous driving software, compiler, runtime, and hardware architecture teams to align long-term roadmap and platform strategy.
Evaluate tradeoffs across latency, throughput, power efficiency, safety, and real-time constraints in production AV systems.
Define benchmarking methodologies and evaluation metrics for next-generation AV AI systems, including robustness, safety, calibration, and edge-case performance.