Gimlet Labs

Gimlet Labs is an applied research lab and inference company building AI systems and infrastructure. Its Gimlet Cloud product provides serverless, managed inference for AI agents, including multi-stage agentic workflows, and runs them across heterogeneous hardware so different parts of a workload can run on the best-suited silicon. The company also develops kforge, a tool that autonomously generates optimized low-level kernels from PyTorch for CUDA, ROCm, and Metal backends. Its customers include frontier labs and companies running large-scale inference workloads, as well as hardware providers seeking deployment.