Neuromorphic AI Kernel Engineer for Edge & Hardware

Intel

Santa Clara (CA)

On-site

USD 180,000 - 260,000

Full time

4 days ago

Be an early applicant

Application generator

Turn this role into an interview — a resume and cover letter built around what this employer wants.

Get past ATS filters

Job summary

Intel’s Neuromorphic Computing Lab in Santa Clara seeks an AI Software Engineer to translate neuromorphic hardware capabilities into high-performance AI software. You will develop specialized kernels, optimize convolutions and transformers, and help shape programming abstractions for the next-generation architecture.

You will work with hardware, compiler, and research teams across the lifecycle from modeling to validation.

Qualifications

PhD with no prior professional experience, a master's degree with 2+ years of relevant experience, or a bachelor's degree with 4+ years of relevant experience in Computer Science, Electrical Engineering, Computer Engineering, Applied Mathematics, Physics, or a related technical field.
4+ years of experience developing, debugging, and delivering maintainable software in Python and either C or C++, including performance-critical or systems-level code.
2+ years of experience implementing and optimizing numerical, machine-learning, or high-performance computing kernels using parallel programming and either an accelerator programming model, such as CUDA, SYCL, or OpenCL , or a domain-specific language, such as Triton .
2+ years of experience developing, training, or evaluating AI algorithms and machine-learning models using a framework

Responsibilities

Design and implement specialized AI kernels for neuromorphic hardware using custom DSLs and accelerator programming models such as CUDA and SYCL and help shape the programming abstractions for Intel's next-generation neuromorphic architecture.
Develop and optimize operations for convolutional networks, transformers, and generative AI workloads by applying tiling, fusion, vectorization, parallelization, layout transformation, buffering, sparsity, and quantization.
Develop performance methodologies and improve kernel behavior in areas such as compute utilization, latency, throughput, memory bandwidth, data movement, synchronization, and scaling, turning architectural insights into measurable workload gains.
Build reference models, numerical validation tools, benchmarks, and analytical or simulation-based performance, power, and area models that guide hardware-software co-design; reconcile discrepancies across models, simulators, emulators, and hardware.
Collaborate with hardware, compiler, runtime, and application engineers to deliver maintainable code, tests, documentation, benchmarks, and performance-regression infrastructure.