Baseten
Software Engineer - Model Performance Systems
$165K–$330K
San Francisco +4 · Hybrid · Full-time · Posted Jan 7, 2026
- Location
- San Francisco
Toronto
New York
Montreal
Seattle
About the job
Build and lead tools and systems for benchmarking, profiling, monitoring, and optimizing LLM inference performance and development environments.
What you'll do
- Automate LLM quality and workload-specific performance benchmarks, and develop tools for evaluation and analysis.
- Build GPU-enabled development environments, performance dashboards, alerts, and CI/CD workflows for performance testing.
- Profile systems, diagnose bottlenecks, and automate the search for optimal model and workload configurations.
What they ask for
- Familiarity with Python.
- Interest in GPU systems, hardware, and data movement across clusters.
- An automation mindset, including stress-testing and fuzzy testing.
- Mathematical curiosity about Transformers, FLOPs, and memory requirements.
Nice to have
- C++ familiarity.
Pay
$165K – $330K (base salary)
Tools
GSM8K, MMLU, InferenceMAX, genai-bench, PyTorch Profiler, NVIDIA Nsight Systems, py-spy, Python
AI focus
Evaluate, benchmark, profile, and optimize LLM inference, including model quality and performance tradeoffs.
Model-written summary · the original posting is authoritative