Teleoperator - Data Annotation for VLA

About Clutterbot

Clutterbot is building robots to maintain, organise, and reset common public and commercial spaces. We are building a hybrid end-to-end architecture where classical and modern methods work together for maximum reliability and robustness.

The role

As a Teleoperator - Data Annotation for Vision-Language-Action (VLA), you will operate

Clutterbot’s robots across lab, real-world, and simulated environments while generating and annotating high-quality data for our VLA models. This is an in-office role based in Bangalore,

combining robotics operations with data annotation and applied AI.

What you'll do

- Teleoperate Clutterbot's clutter-tidying robots in lab settings to perform structured pick-up, sorting, and tidying tasks for demonstration data collection.
- Teleoperate robots in live, real-world food court and cinema hall environments, safely navigating around tables, seating, foot traffic, spills, and varied clutter types (trays, cups, wrappers, popcorn boxes, tissues, etc.).
- Operate robots within simulated software environments to generate synthetic and semi-synthetic demonstration data, stress-test edge cases, and support sim-to-real transfer efforts.
- Maintain smooth, precise, and task-consistent control of robot arms, grippers, and mobility bases across all three settings.
- Follow standardized operating procedures (SOPs) and task scripts designed to produce diverse, high-quality demonstration trajectories.
- Annotate teleoperation sessions and raw sensor data (RGB/depth video, proprioception, language instructions, action trajectories) to produce clean, structured datasets for VLA model training.
- Apply language grounding to robot actions — writing or validating natural-language instructions/descriptions that correspond to demonstrated behaviors,



consistent with VLA data conventions.
- Label and segment task episodes (start/end of sub-tasks, success/failure, contact events, object identities and states) according to annotation guidelines.
- Flag ambiguous, low-quality, or edge-case episodes and route them for review; propose refinements to annotation taxonomies and guidelines as patterns emerge.
- Perform quality checks and audits on annotated data to catch labeling drift, inconsistency, or errors before datasets are used for training.

What you’ll bring

- Experience: 1–5 years in data annotation, robotics operations, quality, or similar roles.
- Education: Graduate in any discipline; technical/engineering background preferred.
- Attention to Detail: Solid attention to detail with a focus on accuracy, consistency, and following defined processes.
- Communication: Good English communication skills and ability to identify and escalate data issues.
- Technical Skills: Comfortable using computers, software applications, and AI/ML or simulation tools.
- Operational Skills: Good spatial awareness and movement control with the ability to operate systems with precision.
- Adaptability: Comfortable working in dynamic, real-world environments and adapting to different tasks
- AI/Robotics: Basic understanding of robotics, AI, or machine learning is a plus.
- Teamwork: Able to work collaboratively with technical and operational teams and contribute constructively to improving processes and data quality.

Why Join Us

- Join a smart, collaborative, world-class team in a fast-paced and flexible environment.
- Opportunity to work at the forefront of innovation in robotics and AI.
- Building a product that people love and use in their daily lives.
- Friendly people, a nice office environment and flexible working.
- Work from our new 23,508 sq ft modern office at Haralur Road, designed for focus, teamwork, and execution.

📌 Teleoperator - Data Annotation for VLA (Karnataka)
🏢 clutterbot
📍 Karnataka