Pulse
Software Engineer, Inference
$150K–$230K
San Francisco · On-site · Full-time · Posted Jul 30, 2025
- Experience
- 3+ years in performance engineering or ML systems.
- Visa sponsorship
- Sponsorship available.
About the job
Build and optimize low-latency, high-throughput inference systems for OCR and multimodal models, including profiling, batching, and autoscaling.
What you'll do
- Build inference services with smart batching and caching; optimize kernels, tokenization, and model graphs.
- Evaluate vLLM, TensorRT LLM, and Triton tradeoffs; implement autoscaling and admission control with SLOs.
- Maintain performance dashboards and capacity plans.
What they ask for
- 3+ years in performance engineering or ML systems.
- Strong Python and exposure to C++ or CUDA.
- Experience with GPU profiling and model serving.
Nice to have
- Experience reducing p95 latency and cost in production ML systems.
- Prior startup or founding experience.
Pay
$150K – $230K USD base salary
Tools
Python, C++, CUDA, vLLM, TensorRT LLM, Triton
AI focus
Inference engineering for OCR and multimodal models, including performance optimization and model serving.
Model-written summary · the original posting is authoritative