The Role

We are hiring a Senior Program Manager – AI Data Operations to own end-to-end execution of our data programs across Physical AI, egocentric video, and Voice AI workstreams. You will be the connective tissue between customers who define data requirements, on-ground collection teams who execute in the field, annotation and labelling pipelines, and internal research and engineering stakeholders who consume the data.

This is a high-ownership, customer-facing, operations-heavy role. You will be as comfortable presenting a delivery report to an enterprise client as you are resolving a data quality blocker with a field collection team at 10pm.

What You Will Do

Program Ownership & Delivery

• Own end-to-end program lifecycle for AI data collection, annotation, and labelling initiatives — from scoping and planning through QA, delivery, and client sign-off.
• Define and maintain program roadmaps, milestones, SLA frameworks, and dependency maps across data collection, annotation vendors, QA teams, and engineering pipelines.
• Drive weekly execution rhythms — standups, delivery reviews, retrospectives, and escalation paths — ensuring predictable throughput and on-time delivery.
• Proactively identify and resolve cross-team blockers, resource gaps, and field-level risks before they impact delivery.

Customer Engagement & Requirements

• Serve as the primary customer-facing point of contact — conducting discovery sessions to understand data requirements, taxonomy definitions, use-case context, and quality benchmarks.
• Translate customer briefs into structured data ops plans: modality specs, annotation schemas, ontology design, collection protocols, and acceptance criteria.
• Provide regular, clear delivery status reports and performance dashboards to enterprise clients and internal leadership.
• Manage customer expectations proactively, flagging risks early and proposing solutions before escalation.

Data Operations & Quality Management

• Manage on-ground data collection teams across diverse environments — wearable cameras, egocentric rigs, audio capture setups, and multi-sensor deployments.
• Oversee annotation and labelling operations across internal and vendor teams using platforms such as Scale AI, Labelbox, CVAT, or equivalent tooling.
• Define and enforce annotation guidelines, quality benchmarks, inter-annotator agreement (IAA) thresholds, and sampling/review workflows.
• Build and maintain operational dashboards tracking throughput, defect rates, annotator productivity, SLA adherence, and cost-per-unit across active programs.
• Drive continuous process improvement in collection protocols, labelling pipelines, and QA frameworks as programs scale.

Cross-Functional Collaboration

• Partner with research and ML engineering teams to understand downstream data consumption requirements — model training pipelines, evaluation benchmarks, and dataset versioning needs.
• Collaborate with hardware and field operations teams to design and iterate on data collection protocols for real-world deployment environments.
• Contribute to building scalable, repeatable processes and playbooks that grow with the Physical AI data program.

What We Are Looking For

Must-Have

• 5–9 years of program or project management experience, with at least 3 years managing AI data operations — data collection, annotation, labelling, or data pipeline programs at scale.
• Hands-on experience with annotation platforms and tooling: Scale AI, Labelbox, CVAT, Supervisely, Encord, or equivalent.
• Working knowledge of Physical AI, egocentric video, or Voice AI data taxonomy — including scene understanding, action segmentation, affordance labelling, bounding box/polygon/keypoint annotation, intent classification, utterance tagging, and ontology/schema design.
• Proven track record as a customer-facing program manager — conducting requirements discovery, managing client relationships, and delivering structured status reporting.
• Strong analytical skills: comfortable with Excel/Sheets, SQL basics, and operational dashboards for throughput, quality, and SLA tracking.
• Experience managing distributed, on-ground field teams and/or annotation vendors across multiple concurrent programs.
• Exceptional written and verbal communication — able to simplify complex data ops problems for both technical and non-technical audiences.

Strong Plus

• Direct experience with egocentric video data programs — wearable camera rigs, first-person perspective annotation, gaze/intent/action labelling for embodied AI or AR/VR.
• Experience with Voice AI or conversational AI data: utterance collection, transcription, speaker diarisation, wake word labelling, or NLU annotation workflows.
• Familiarity with robotics perception data — LiDAR point cloud annotation, depth sensing, multi-modal sensor fusion datasets.
• Understanding of data quality frameworks: inter-annotator agreement (IAA/Kappa), gold standard benchmarking, edge case curation, and dataset versioning.
• Exposure to ML training pipelines and how dataset structure, volume, and quality affect model performance.
• Experience with tools like Jira, Confluence, Notion, Airtable, or equivalent for program tracking and documentation.
• Bachelor's or Master's degree in engineering, computer science, data science, or a related field.

Who You Are

• You thrive in ambiguity and build structure where none exists — turning loosely defined customer asks into airtight execution plans.
• You are a relentless problem-solver who can fix a field collection blocker as comfortably as you can present a program review to a C-suite stakeholder.
• You are deeply process-oriented and analytically driven — you measure everything, and you use data to make decisions and catch problems early.
• You care about the quality and integrity of data, and you understand that the models are only as good as what goes into them.
• You own outcomes, not just tasks.