Tensortune

Tensortune provides AI inference performance engineering services to make AI models faster and cheaper. They specialize in low-level optimization of LLM, VLM, speech, computer vision, and forecasting models, down to the kernel level, and serve them with low latency to many concurrent users. Their services include runtime optimization, low-latency multi-user streaming server development, profiling and bottleneck analysis, fine-tuning, and deployment strategy.