Serve Robotics is actively changing how items are moved via sidewalk robots which provide deliveries without street congestion. They are a team of tech industry veterans in software, hardware, and design are pooling their skills to build the future; their team is agile, diverse, and driven.
Design and implement production-grade auto-labeling pipelines that generate 3D and 4D annotations from multi-modal robot data at scale.
Develop data-centric learning workflows that connect auto-label outputs, Serve’s dataset infrastructure, and continuous E2E model training and evaluation pipelines.
Lead initiatives in self-training, weak supervision, and simulation-to-real adaptation to reduce manual labeling dependency and accelerate model iteration cycles.