
Data & Model Quality Engineer - Foundation Model Team
TELEXISTENCE Inc.16 hours ago
Tokyo, JapanMid Level
Responsibilities
- Run model training across hyperparameter configurations and select stable checkpoints for evaluation.
- Define evaluation frameworks, metrics, and KPIs for models and datasets, including success rates, failure modes, and variance across runs.
- Use evaluation results to determine future data-collection priorities and maintain definitions of good and bad data.
- Build tools and methods to detect, flag, and filter problematic multimodal data.
- Create collection guidelines, train operator leads, and monitor adherence to quality standards.
- Build or contribute to infrastructure for automated model and data-quality evaluation and pipelines that transform raw multimodal data into training and evaluation datasets.
- Collaborate with Hardware, Operations, and System Engineering teams and incorporate advances in robot learning, data curation, and evaluation.
Requirements
- Experience training robotics-control models, including hyperparameter tuning.
- Ability to understand model architectures and use that understanding to tune hyperparameters and improve performance.
- Experience reviewing data, identifying problematic content, and improving data quality, ideally for multimodal data including images, video, trajectories, and sensor logs.
- Strong Python programming skills; experience with other languages is a plus.
- Experience working in a fast-paced, high-output environment.
- Strong English communication and ability to translate technical findings into practical instructions for operators.
Tech Stack
Categories
ML EngineeringRobotics
About TELEXISTENCE Inc.
TELEXISTENCE Inc. builds and deploys industrial robots and a cloud software platform for logistics and retail operations, including manipulator systems and automated forklifts (AGFs). Its products focus on teleoperation and AI-driven autonomy for tasks like picking, motion/path planning, and system integration at production and warehouse sites. Founded in 2017 and headquartered in Ota-ku, Tokyo, the privately held company serves retailers and logistics operators in Japan and beyond.