Teams at SpAItial AI

Own recurring business operations, including finance cycles, payroll, onboarding, compliance, vendors, workplace management, process documentation, and reporting. Partner with the COO, accountants, advisors, employees, suppliers, and German colleagues to maintain accurate finances, reliable workflows, regulatory compliance, and effective office operations at a fast-moving AI startup.

Own end-to-end training of vision-language-action and world-action robot policies, from dataset curation and model architecture through distributed training, evaluation, and deployment. Responsibilities include adapting VLM backbones for control, designing action representations, reducing the sim-to-real gap, building action-conditioned world models, and applying supervised fine-tuning and reinforcement learning. The role also contributes to technical direction for embodied AI research.

Research Scientist developing robust 3D reconstruction systems using multi-view geometry, SfM, SLAM, bundle adjustment, deep multi-view stereo, learned matching, and monocular depth. Responsibilities include designing pose estimators and reconstructors, scaling methods to real-world datasets, evaluating accuracy and quality, integrating state-of-the-art models, and collaborating on production deployment. Requires advanced computer vision research expertise, Python and PyTorch proficiency, and practical experience with COLMAP, ORB-SLAM, and Ceres.