Policy & Quality Specialist- ML Perception Data

Job description

Job responsibilities

• Translate requirements to policies: Collaborate directly with MTV-based Perception Engineers/ ML model owners to understand their specific data goals, translate their ambiguous machine learning requirements into…
• Drive queue readiness & golden datasets: Speed up the initial labeling queue setup process by executing rapid policy iterations and hand-crafting golden datasets (small-scale baseline datasets of 10s of examples) to…
• Direct vendor teams: Provide technical guidance and operational direction to vendor labeling experts to enable rapid policy setup and ensure that the ~10 active labeling queues under your purview run smoothly and meet…
• Address edge cases & regional nuances: Provide critical, detailed inputs on long-tail edge cases and coordinate with regional country specialists to ensure country-specific driving rules and local nuances are…
• Enable quality and process improvements: Monitor labeling pipelines, conduct targeted technical analyses to identify data quality trends, and build/maintain automated data analysis tools to proactively identify…
• Facilitate cross-functional knowledge sharing: Act as the primary technical interface between requesters and operations, ensuring on-ground dissipation of policies, managing policy amendments, and resolving complex…

Job requirements

• 4-5+ years of experience in data analysis, operations, or program management with a focus on machine learning data annotation, taxonomy design, or human-in-the-loop workflows.
• Operational project management: Demonstrated ability to work independently on operational workflows and successfully project manage small sub-working groups or vendor squads.
• Core ML data lifecycle understanding: Practical knowledge of dataset curation, labeling pipelines, data quality control metrics, and baseline model evaluation concepts.
• Analytical aptitude: Experience conducting technical data analyses using pre-established tools (or building simple automation scripts) to diagnose pipeline issues, track vendor quality, and generate actionable insights.
• Adaptable & detail-oriented: Comfort working within a dynamic environment, translating vague technical needs into clear documentation, and maintaining a high standard of attention to detail.