Head of ML Data Quality & Labeling (Remote)

Coalition, Inc.; United States; Remote; USD 87,515 - 132,801; Full time

Benefits offered by this job

100% medical, dental, and vision

Flexible PTO

Annual home office stipend and WeWork

Mental & physical health programs

Competitive compensation and growth

Job summary

Coalition is a remote-first organization focused on preemptive cyber risk coverage. We seek a data-labeling leader who will own data quality, task design, and annotation standards to ensure production-ready datasets for ML models.

You will manage Label Studio configurations, align with ML and product teams, and oversee external labeling vendors to meet quality and throughput targets, reporting to the Chief Product Officer.

Qualifications

• 5+ years in ML data operations, data labeling, or a related field (ML engineering, data science, or data engineering with heavy labeling exposure)
• Deep understanding of annotation quality frameworks: IAA, consensus labeling, gold standard evaluation, error taxonomy, and calibration workflows
• Direct experience managing labeling platforms (Label Studio strongly preferred; Scale AI, Labelbox, Prodigy, or similar acceptable)
• Track record managing outsourced labeling vendors or BPOs for ML data production
• Familiarity with common ML labeling tasks: text classification, NER, document extraction, intent detection
• Comfortable working in Python and SQL; bonus if you’ve built tooling around labeling workflows or quality measurement
• Strong opinions on what makes labeled data good or bad, and the willingness to push back when it’s bad
• Experience in insurance, cybersecurity, or fintech is a plus but not required

Responsibilities

• Labeling quality & methodology: Define annotation guidelines, taxonomies, and edge-case protocols for each labeling program. Establish gold standard datasets, inter-annotator agreement targets, and audit sampling processes. Identify and remediate mislabeled data in existing datasets.
• Platform & tooling: Serve as the primary user and requirements driver for Label Studio — defining project configuration needs, workflow designs, pre-labeling pipeline requirements, and integration points with ML infrastructure. Partner with the data engineering team that builds and maintains the platform.
• Cross-functional partnership: Work with ML engineers, data scientists, and product managers to translate model requirements into well-structured labeling tasks. Challenge teams on task design when labeling instructions are ambiguous or likely to produce unreliable labels.
• Vendor management: Source, onboard, and manage external labeling vendors and BPOs in coordination with Coalition’s operations team. Set quality SLAs, run calibration sessions, and manage feedback loops to labelers. Hold vendors accountable to accuracy, not just throughput.
• Measurement & improvement: Define and track operational metrics — label accuracy, IAA scores, cost per label, turnaround time — and use them to drive continuous improvement. Identify opportunities for active learning, model-assisted labeling, and pre-annotation to reduce cost without sacrificing quality.

Job description

Coalition is a remote-first organization focused on preemptive cyber risk coverage. We seek a data-labeling leader who will own data quality, task design, and annotation standards to ensure production-ready datasets for ML models.

You will manage Label Studio configurations, align with ML and product teams, and oversee external labeling vendors to meet quality and throughput targets, reporting to the Chief Product Officer.