Speaker: Ben Segal
Software Engineer @Parasail
Ben Segal works on inference performance at Parasail, an inference cloud that serves open-weight models on its own distributed GPU fleet. He focuses on the measurement and automation behind serving frontier models efficiently, and on the infrastructure that lets engineers and agents experiment on real hardware and real traffic without putting production at risk. Before Parasail, Ben spent about eight years at Swift Navigation, most recently as Senior Director of Engineering running Skylark, a global real-time service that corrects GPS to centimeter accuracy. He was also CTO of Navmatic, a micromobility safety startup acquired by Superpedestrian.
Session
Efficient GPU Agents: Infra for Autonomous LLM Perf Optimization
Parasail serves open-weight LLMs on distributed clusters of GPUs managed by Kubernetes. The typical performance optimization loop is to reproduce a problem or measure a regression, iterate on code or config, set up and run benchmarks, and decide whether the change is worth shipping.