AWS Trainium Kernel Engineer

This is a fully remote position, open to applicants in New York.

📋 Description

• Evaluate kernels developed utilizing the Neuron Kernel Interface (NKI)

• Determine if implementations effectively target AWS Trainium and Inferentia2 hardware

• Examine low-level computation patterns for technical accuracy

• Review the migration of existing CUDA kernels to NKI

• Assess tile decomposition, computational strategies, and partitioning choices

• Analyze the use of SBUF, PSUM, and HBM across Trainium memory hierarchies

• Evaluate DMA orchestration, data movement, sequencing, stalls, and synchronization

• Review Trainium-specific optimization strategies and identify performance bottlenecks

• Assess NeuronCore pipeline utilization, tensor-engine throughput, and memory-bandwidth behavior

• Evaluate numerical consistency between GPU and Trainium implementations

• Review precision choices, mixed-precision behavior, numerical stability, and supported formats

• Evaluate implementations using the AWS Neuron SDK, compiler, and NKI kernel libraries

• Analyze benchmark results for Trn1 or Trn2 workloads and evaluate experimental methodology

• Assess assigned kernel-development tasks against established technical criteria

• Provide clear written, rubric-based technical feedback supported by implementation, profiling, or numerical evidence

⛳️ Requirements

• 2+ years of hands-on experience in developing or optimizing kernels using the Neuron Kernel Interface (NKI)

• Professional experience targeting AWS Trainium or Inferentia2 hardware

• Strong understanding of tile-based computation

• Deep familiarity with SBUF, PSUM, and HBM memory management

• Strong knowledge of partition-dimension constraints and DMA orchestration

• Proven experience evaluating or conducting CUDA-to-NKI migrations

• Familiarity with Trainium-specific performance profiling

• Experience in assessing NeuronCore pipeline utilization, tensor-engine throughput, and memory-bandwidth bottlenecks

• Strong understanding of cross-platform numerical correctness and mixed-precision behavior

• Excellent written communication skills and ability to provide precise technical feedback

• Direct experience with the AWS Neuron SDK, Neuron Compiler internals, or NKI kernel libraries is preferred

• Prior experience in CUDA or Triton kernel development is advantageous

• Familiarity with NeuronCore-v2 architecture and supported numerical formats is preferred

• Experience benchmarking ML workloads on Trn1 or Trn2 instances is advantageous

• H1-B and STEM OPT support is not available for this engagement

🏝️ Benefits

• Part-time independent contractor engagement

• Fully remote position within the United States

• Flexible scheduling based on project requirements

• Projects may be extended, shortened, or concluded based on project needs and performance