Lead Data Collection Ops for Robotics Lab

Dover; Singapore; Hybrid; SGD 60,000 - 90,000; Full time

Job summary

UMA is seeking a Lead Operator - Data Collection to own the pipeline feeding our models. You’ll work with the research and hardware teams to define what a good episode looks like, debug the stack when it stops, and turn operators into an operation that researchers can rely on.

Responsibilities include owning the collection plan end to end, enforcing quality standards, maintaining dataset hygiene, and providing first-line technical support for stations and robots.

Qualifications

• 3+ years hands-on in lab ops, robotics, hardware test, field ops, or data ops.
• Proficiency with Python, CLI (running/editing scripts), and Git.
• Hands-on team leadership on the floor.
• Fluency in French and English.

Responsibilities

• Own the collection plan end to end: translate research needs into weekly plans, task definitions, and targets.
• Gather data with the team to understand processes and manage peak workloads.
• Enforce quality standards: define, inspect, validate, and re-collect episodes as needed.
• Maintain dataset hygiene, labeling, structure, traceability, and coverage.
• Provide first-line technical support for stations, control interfaces, and robots.
• Shape inspection and correction tools by giving product feedback.
• Partner with hardware and software teams to resolve recurring failures.
• Author scalable SOPs, training materials, and operator documentation.
• Share recurring metrics on episode counts, acceptance rates, coverage, uptime, and saved researcher hours.
• Ensure safety on the floor and keep the workspace clean and organized.

Skills

Lab operations

Robotics

Hardware testing

Field operations

Data operations

Python

CLI

Git

Team leadership

Bilingual (Fr-Eng)

Tools

ROS

Hugging Face

LeRobot

QA/annotation pipelines

Job description

UMA is seeking a Lead Operator - Data Collection to own the pipeline feeding our models. You’ll work with the research and hardware teams to define what a good episode looks like, debug the stack when it stops, and turn operators into an operation that researchers can rely on.

Responsibilities include owning the collection plan end to end, enforcing quality standards, maintaining dataset hygiene, and providing first-line technical support for stations and robots.