Teams at Mecka AI

Own end-to-end scientific data acquisition programs for AI-lab customers: scope projects, recruit and manage experts, design protocols, define quality and metrics, run pilots to production, and coordinate cross-functional teams to deliver model-useful datasets.

Lead development of simulation environments, retarget human demonstrations to robot trajectories, train and evaluate policies using imitation learning and RL, and drive sim-to-real transfer with metrics and domain-randomization methods.

Build performant web viewers/players and SDKs for egocentric robotics data. Design multi-stream playback, timeline scrubbing, sync, and APIs; own interface contracts across hardware, data, and consumers; provide tooling and reference implementations for internal and external developers.