Senior LLM Inference Performance Engineer
1 day ago
Winnipeg MB, Winnipeg Census Division, MB; Manitoba, Canada
Aptino, Inc.
Full-time
Free with email or Google
Save this job and keep your search organized
Create a free account to save jobs, create alerts and return to this listing from your dashboard.
Free with email or Google
By continuing, you agree to our Terms & Privacy Policy.
Optimize large language model inference with Python and PyTorch across AWS Trainium platforms, including Trainium2 and Trainium3. Improve production AI performance through hands-on accelerator optimization.
As a Senior Machine Learning Systems Engineer, you will improve inference latency, throughput, scalability, and serving costs. The work includes LLM inference frameworks, AWS Neuron SDK, vLLM on Neuron, and production workload optimization. You will evaluate advanced techniques for Qwen-family models.
Key Responsibilities
Optimize decode latency, throughput, and LLM serving costs Work with Trainium2, Trainium3, Neuron SDK, and vLLM Evaluate speculative decoding and Multi-Token Prediction for Qwen models Profile and benchmark GPU and Trainium production workloads Optimize prefill, decode, batching, KV-cache, and parallelism
Requirements:
Five-plus years in Machine Learning or Systems Engineering Two-plus years with production LLM inference or serving Strong Python and PyTorch programming experience Experience with vLLM, TensorRT-LLM, SGLang, or Triton Understanding of prefill, decode, batching, and KV-cache management Apply inference performance engineering, Python, PyTorch, and accelerator optimization across demanding production LLM workloads. #J-18808-Ljbffr
Key Responsibilities
Optimize decode latency, throughput, and LLM serving costs Work with Trainium2, Trainium3, Neuron SDK, and vLLM Evaluate speculative decoding and Multi-Token Prediction for Qwen models Profile and benchmark GPU and Trainium production workloads Optimize prefill, decode, batching, KV-cache, and parallelism
Requirements:
Five-plus years in Machine Learning or Systems Engineering Two-plus years with production LLM inference or serving Strong Python and PyTorch programming experience Experience with vLLM, TensorRT-LLM, SGLang, or Triton Understanding of prefill, decode, batching, and KV-cache management Apply inference performance engineering, Python, PyTorch, and accelerator optimization across demanding production LLM workloads. #J-18808-Ljbffr