Research Engineer-Model Efficiency
Core
Implement and adapt model-efficiency methods for LLM inference to improve serving cost and speed while preserving quality.
Role type
Research Engineer specializing in model efficiency and LLM inference optimization
Builds
Optimized LLM serving systems and efficient inference pipelines
Domain
Artificial Intelligence / Large Language Models / Model Efficiency
Deliverable
production ML models
Required skills
LLM inference, model optimization, Python, PyTorch, transformer internals, quantization, sparsity, pruning, speculative decoding, MTP, serving-time attention, KV-cache methods, rigorous evaluation
Preferred skills
C++, CUDA, Triton, vLLM, SGLang, TensorRT-LLM, production experience with efficiency methods, publications at top-tier venues, open-source contributions
Technologies
PyTorch, C++, CUDA, Triton, vLLM, SGLang, TensorRT-LLM
Responsibilities
Implement and adapt published model-efficiency methods for models and hardware; Develop new optimizations for efficient LLM serving; Build rigorous evaluations using task-level metrics and controlled comparisons; Assess techniques including quantization, sparsity, pruning, and speculative decoding
Seniority
Mid-to-Senior, hands-on IC
