AI Systems Architect (SoC) at D24 Search AI Systems Architect (SoC) at D24 Search in Campbell, California Posted in about 2 hours ago.
Type:
full-time
Job Description:
AI Systems Architect (SoC) Santa Clara, CA (Hybrid) Up to $325K + Equity My client is building next-gen compute platforms for physical and embodied AI
- the next wave after LLMs.
They're uniquely positioned as the only company doing custom silicon for robotics and autonomous systems. Join a well-funded startup with a stellar board including Intel's CEO and Forbes Midas List investors. The team has already delivered 33+ million chips and brings proven execution to this massive emerging market. This is a founding AI systems architect role with massive influence
- you'll be the first hire in this function reporting to the CDO.
High visibility position shaping compute platforms for the future of intelligent machines. Role Overview We are looking for an AI Systems Architect who sits at the intersection of machine learning models and hardware architecture . In this role, you will work across model development, algorithms, and hardware architecture to identify and shape the AI workloads that will define the next generation of compute platforms. You will evaluate emerging model architectures, understand their training and inference characteristics, and help translate them into efficient hardware implementations. This role requires someone who is comfortable moving between mathematical models, software frameworks, and hardware architecture , and who enjoys solving problems at the boundary between disciplines.
Requirements:
5
- 7 years of experience in mapping AI models onto hardware Mapped AI models (e. g. , Transformers) to hardware/silicon Optimized models for performance/ efficiency
- quantization, pruning, efficiency improvements, memory footprint reduction Experience at a top AI lab or hardware company (e. g. , Nvidia, Meta) PhD with hands
- on, real
- world implementation experience is a great plus Deep knowledge of Transformer model architecture VLA (Vision Language Action Models), VLM (Vision Language Models), ViT (Vision Transformers)
- some familiarity is required NPU/SOC design understanding
- mapping models onto neural processing units; understanding chip
- level hardware constraints Experience with ML frameworks like PyTorch, JAX, or TensorFlow. High influence, high visibility mindset
- will be first robotics hire and must be comfortable shaping direction without a large team around them Ability to work hybrid or onsite in the Bay Area (Santa Clara) Responsibilities Analyze modern AI model architectures including transformers and emerging alternatives to understand their computational and system requirements.
Evaluate model suitability for different application domains and help determine which model architectures are best matched to specific workloads. Work closely with hardware architects to translate model requirements into efficient silicon implementations. Identify opportunities to modify or optimize models, algorithms, or dataflows to improve performance, efficiency, or scalability in hardware. Understand training and inference pipelines for modern AI models and identify implications for hardware design. Develop performance and efficiency models to guide architectural decisions. Collaborate with software and hardware teams to ensure that models can be deployed efficiently on new compute platforms. Stay current with emerging AI model research and identify opportunities where new architectures may benefit from specialized hardware. Required Qualifications Strong understanding of modern machine learning architectures, including transformers and related model families. Solid grounding in the mathematical foundations of machine learning, optimization, and deep learning algorithms. Experience working with ML frameworks such as PyTorch, JAX, or TensorFlow. Ability to analyze model computation graphs and translate them into efficient dataflows and compute patterns. Experience with AI model training and inference workflows. Strong systems thinking and ability to work across software, algorithms, and hardware. Preferred Qualifications Experience with hardware-software co-design or AI accelerator architecture. Familiarity with model optimization techniques such as quantization, sparsity, pruning, or distillation. Experience implementing or optimizing AI models for specialized hardware platforms. Exposure to emerging AI model architectures beyond transformers (e.g., world models, continuous-time networks, reinforcement learning systems). Experience working with robotics, autonomous systems, or embodied AI applications. Why Join? You will work on problems at the frontier of AI systems and compute architecture , helping shape how the next generation of AI models-especially those operating in the physical world
- is translated into silicon.
This is an opportunity to work with a small team of experienced builders who have designed and delivered some of the industry's most widely deployed compute platforms. If you are excited about bridging the worlds of machine learning models and hardware architecture , we would love to hear from you.
