Speech AI Data Operations Lead

Commotion; New York (NY); Hybrid; USD 120,000 - 180,000; Full time

Job summary

Commotion seeks a data operations lead to build and own the data engine behind our Voice AI products in New York. You will translate research needs into scalable data collection, annotations, and licensing workflows across multilingual datasets.

You will manage external vendors, design QA frameworks, and drive end-to-end governance from sourcing to production-ready datasets for ASR, TTS, and conversational models.

Qualifications

• Experience in AI data operations and large-scale data collection.
• Proven ability to manage external vendors and multi-stakeholder projects.
• Strong negotiation and communication skills, and cross-functional collaboration.

Responsibilities

• Define scalable data strategies for multilingual speech data across accents and domains.
• Identify, onboard, and manage vendors, studios, and freelancers with SLAs.
• Design annotation workflows, guidelines, and QA frameworks for transcription and labeling.
• Improve quality and turnaround of data products continuously.
• Coordinate voice talent recordings and manage consent, licensing, and metadata.
• Define dataset quality metrics, audits, and documentation governance.
• Own timelines and budgets across research, engineering, product, procurement, and legal.

Skills

Data operations

Vendor management

Project management

Negotiation

Communication

Tools

Label Studio

SQL

Python

Job description

You'll build and own the data engine behind our Voice AI products. Sitting at the intersection of AI research, engineering, and operations, you'll run the full speech-data lifecycle from defining requirements to sourcing vendors, managing annotation, procuring voices, and delivering production-ready datasets for training and evaluating ASR, TTS, and Conversational AI models. This is a build-from-scratch role for someone who likes turning messy operational problems into scalable processes.

Responsibilities:

• Data strategy: Translate research and product needs into scalable plans for collecting multilingual speech data across accents, demographics, domains, and recording conditions.
• Sourcing and procurement: Identify, onboard, and manage vendors, studios, voice agencies, crowdsourcing platforms, and freelancers. Negotiate pricing, contracts, and SLAs, and track performance against quality and delivery metrics.
• Annotation operations: Design annotation workflows, guidelines, SOPs, and QA frameworks for transcription, speaker/intent labeling, translation, and evaluation.
• Continuously improve quality and turnaround.
• Voice talent and licensing: Recruit voice artists, coordinate recordings, and manage consent, licensing, usage rights, and voice metadata.
• Quality and governance: Define dataset quality metrics, run audits and root-cause analysis, and maintain versioning, documentation, traceability, and privacy/licensing compliance.
• Program management: Own timelines, budgets, and execution across research, engineering, product, procurement, finance, and legal. Flag risks early and report progress via operational dashboards.

Requirements:

• 3+ years in AI data operations, annotation operations, data collection, or program management.
• Proven experience managing external vendors and large-scale annotation projects.
• Strong project management, stakeholder management, negotiation, and communication skills.
• A track record of building operational processes and owning cross-functional initiatives independently.
• Nice to have: Experience with ASR / TTS / Conversational AI, multilingual or Indian-language datasets, human evaluation pipelines, voice recording operations, or data privacy and licensing. Basic SQL/Python and familiarity with Label Studio (or similar) are a plus.