Teams at GRAM (gramcorporation.com)

Build infrastructure for reproducible robot training and evaluation datasets, including multimodal ingestion, temporal alignment, provenance, quality controls, storage, versioning, replay, and reliable delivery. Own schemas, lineage, observability, failure recovery, safe backfills, and dataset reproducibility across video, sensor, command, intervention, outcome, and machine-health data.

Own the architecture and integration of onboard sensing, compute, communications, power, data interfaces, timing, calibration, and software pathways for autonomous insectoid robots. Select and integrate sensors, embedded boards, compute modules, harnesses, and middleware; define interface contracts; characterize timing and latency; and diagnose hardware-software faults involving communications, power, thermal behavior, calibration, and mechanical alignment. Write C++ or Python integration code and maintain compatibility and calibration records.

Lead the architecture and hands-on development of manipulation systems for physical insectoid robots. Design contact-rich behaviors, planning, learned-policy integration, force control, monitoring, recovery, and real-time execution in C++ and Python. Establish experiments, metrics, regression tests, and failure-analysis processes that improve repeated physical-task performance. Mentor engineers, lead cross-stack reviews, and make technical decisions across sensing, control, actuation, interfaces, and mechanical constraints.