Software Engineer - Model Performance Systems
Baseten · San Francisco
About this role
**Software Engineer - Model Performance Systems** **About Baseten** Baseten powers mission-critical inference for leading AI companies like Cursor, Notion, OpenEvidence, Abridge, Clay, Gamma, and Writer. Recently raised $1.5B Series F. **The Role** A specialized, high-impact position at the intersection of high-performance computing (HPC) and Large Language Model (LLM) engineering. You'll build automated performance evaluation tools and define the roadmap for next-generation AI infrastructure. **Key Responsibilities** • Benchmarking: Evaluate and automate LLM quality benchmarks (GSM8K, MMLU) and custom performance suites • DevEx Improvement: Develop GPU-enabled development environments for model experimentation • Tool Development: Build and contribute to open-source tools (InferenceMAX, genai-bench) • System Profiling: Use PyTorch Profiler, NVIDIA Nsight Systems, and py-spy to identify bottlenecks • Monitoring & Observability: Develop real-time dashboards and performance alerts • Continuous Integration: Automate performance testing and release workflows • Optimization Automation: Identify optimal configurations (latency vs. cost vs. quality) **What We're Looking For** • Deep systems & hardware knowledge (GPU memory, InfiniBand, cluster networking) • Strong automation mindset and stress-testing expertise • Mathematical curiosity about Transformers, FLOPs, and memory requirements • Proficiency in Python; NVIDIA software stack expertise; C++ a plus • Mid-senior level with strong communication and mentoring abilities **Benefits** • Competitive compensation with meaningful equity • 100% medical, dental, vision coverage (U.S.) • Flexible PTO + company Winter Break • Paid parental leave & fertility stipend • 401(k) facilitation (U.S.) • Exposure to diverse ML startup ecosystem *Equal Opportunity Employer*
Listing freshness
CronJobs last confirmed this listing 1h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.