Senior Performance Analyst, Inference

Tasks

• Build and maintain competitive pricing models
• Build reproducible inference benchmarks
• Create deal specific competitive analyses for enterprise prospects
• Design standardized benchmark suites for inference workloads
• Measure tokens per second time to first token latency under concurrency TCO
• Monitor industry announcements and pricing changes
• Partner with sales to recommend pricing for contracts
• Synthesize findings into competitive briefs for sales and product
• Track third party benchmarking sources and validate measurements

Perks/Benefits

• N/A

Skills/Tech-stack

Attention Mechanism | CUDA | Flash Attention | GPU kernel optimization | KV cache | Kernel optimization | Quantization | SGLang | TensorRT | TensorRT-LLM | Transformer | Triton | VLLM