🛠 Tech Stack

💼 About This Role

You'll make our inference stack as fast and efficient as possible, working from CUDA kernels to serving frameworks to eliminate bottlenecks. Your north star is inference performance: latency, throughput, and cost efficiency. You'll have autonomy, a large compute budget, and technical support to push limits.

🎯 What You'll Do

• Implement optimization techniques like quantization and speculative decoding
• Deep dive into inference frameworks to debug and improve performance
• Profile and optimize CUDA kernels and GPU utilization
• Build tooling and benchmarks to track inference performance

📋 Requirements

• 2+ years of experience in ML systems or inference optimization
• Strong proficiency in Python and familiarity with C++
• Hands-on experience with LLM inference frameworks (vLLM, SGLang, TensorRT-LLM)
• Deep understanding of GPU architecture and profiling GPU workloads

✨ Nice to Have

• Experience with CUDA programming
• Familiarity with serving non-LLM models (TTS, vision, embeddings)
• Contributions to open-source inference frameworks

🎁 Benefits & Perks

• 💰 Competitive base salary $220,000 - $320,000
• 📈 Equity in high-growth startup
• 🏥 Comprehensive benefits
• 🏢 In-person collaboration in downtown SF

📨 Hiring Process

[email protected]