Staff, Robotics ML/Data Engineer

About the role

Persona AI is building humanoid robots for heavy‑industry environments such as shipyards, steel mills, fabrication facilities, and offshore platforms. The company needs massive, high‑quality multimodal datasets—egocentric video combined with dense sensor streams (IMU, haptics, kinematics, force profiles)—to train its robots. As a Staff, Robotics ML/Data Engineer you will architect and scale the infrastructure that transforms raw, unstructured recordings into training‑grade data, directly influencing how quickly the foundation models learn dexterous manipulation.

Responsibilities

• Force Analysis & Hidden State Inference: design cross‑modal validation systems that verify consistency among video, proprioception, force/haptic signals, and language annotations.
• Kinematic Retargeting & Alignment: orchestrate hand‑tracking, segmentation, depth estimation, 3D reconstruction, and pose‑tracking modules; retarget human demonstrations into robot trajectories; run simulation‑in‑the‑loop validation for physical plausibility.
• Advanced Data Augmentation: implement spatial, temporal, synthetic‑viewpoint, and sensor‑noise augmentations to expand expert trajectories and improve model robustness.
• Teleoperation Synchronization: create unified state‑action representations across differing robot embodiments, coordinate frames, rotation conventions, and sampling rates, with per‑dimension validity masking and normalization.
• Close the loop with data consumers: build tooling for researchers to query, visualize, and audit datasets (clip browsers, trajectory viewers, annotation review UIs) and translate model‑failure analyses into new curation rules.
• Multimodal Data Pipelines: architect end‑to‑end ingestion pipelines that convert raw recordings into indexed, queryable, training‑ready datasets, including temporal segmentation, metadata extraction, scene‑graph generation, embedding‑based retrieval, and language annotation workflows.

Qualifications

• Education: M.S. or Ph.D. in Computer Science, Data Engineering, Machine Learning, Robotics, or a related field.

Skills

• Programming & ML Frameworks: deep expertise in Python and extensive experience with PyTorch, especially custom dataloaders for multimodal datasets.
• Force & Time‑Series Data Processing: experience analyzing and aligning complex force‑torque, load‑cell, or tactile array data with visual frames.
• Video Processing: mastery of video pipelines and libraries (OpenCV, FFmpeg, Decord) and handling terabyte‑scale video I/O.
• 3D Geometry & Robotics Data: solid knowledge of coordinate frames, transforms, rotation representations, camera intrinsics/extrinsics, forward/inverse kinematics, and URDF.
• Data Augmentation: proven ability to implement programmatic and generative augmentation techniques for computer‑vision and time‑series data.

Bonus Skills

• Experience with NVIDIA’s robotic software stack (Open X‑Embodiment, DROID, AgiBot World, EgoDex, or similar).
• Familiarity with modern perception toolboxes: segmentation (SAM‑family), monocular depth, hand/body pose estimation (MANO/SMPL), 6‑DoF object pose tracking, point tracking.
• Experience with distributed data‑processing systems (Ray, Apache Spark) for cluster computing.
• Background in generating or using synthetic robotic data via simulation (Omniverse, MuJoCo).
• Experience integrating spatial awareness or tactile data representations (e.g., Fourier encoding) into visual pipelines.

Benefits

• Competitive compensation with performance‑based bonus.
• 99% employer‑covered medical benefits.
• Early‑stage equity participation.
• Competitive paid time off.
• Company‑wide paid winter break (December 24 – January 2).