Physical AI for autonomous warehouse robots handling pick, pack, and sort tasks
Dexterity builds full-stack robot control software for warehouse and logistics automation, running deep learning (PyTorch, TensorFlow) and physics-based modeling on C++ and Python cores deployed across AWS/GCP/Azure. The engineering-heavy hiring mix—18 engineers concentrated at senior/mid levels with focus on packing algorithms, sensor fusion, and low-latency inference—reflects active work on production deployment challenges: the company is scaling ML pipelines while managing field reliability and supply-chain resilience across live warehouse environments.
Dexterity develops full-stack dexterity solutions for robotic systems in logistics and warehousing. The product stack spans real-time perception (sensor fusion, state estimation), physics-aware ML models for manipulation planning, and production-grade control software deployed in live customer environments. Core projects include improving packing density algorithms, training data collection at scale, and lifecycle stewardship of deployed systems. The team operates across United States and Canada, with hiring velocity decelerating—a pattern typical of companies moving from R&D toward operational maturity.
C++, Python, PyTorch, TensorFlow for ML; Kubernetes, Docker for infrastructure; AWS, GCP, Azure for cloud; Solidworks and NX for CAD; Redis and gRPC for real-time communication.
Field deployment reliability, low-latency inference at scale, packing algorithm optimization for denser structures, scaling ML pipelines, and maintaining security guidelines for production robotics systems.
Other companies in the same industry, closest in size