Technical Program Manager, Data Collection and Model Evaluation (Fixed-Term Contract), DeepMind

• Manage complex human data collection and model evaluation programs spanning pre-training, SFT, and benchmark evaluations across multimodal domains from requirement intake through final data handoff.
• Partner with research and engineering to translate modeling goals into concrete data specifications, annotation guidelines, and evaluation rubrics, while driving workflow migrations onto internal platforms.
• Coordinate technical dependencies for novel data collection environments (e.g., virtualized OS, multimodal capture), resolving tooling bottlenecks and ingestion issues.
• Oversee external data service providers and rater pools, managing SOWs, capacity, throughput SLAs, and quality audits to ensure datasets meet research quality bars.
• Establish program tracking and operational metrics; clearly communicate progress, risk mitigation plans, and resource status to cross-functional partners and leadership.

Minimum qualifications:

• Bachelor’s degree in a technical, quantitative, or operational field, or equivalent practical experience.
• 5 years of experience in technical program management, data operations, or AI/ML program management.

Preferred qualifications:

• Manage end-to-end human data collection, annotation, and evaluation pipelines (SFT, RLHF) for large-scale foundation models.
• Oversee third-party vendors, SOWs, and rater quality, partnering with engineering to design and launch workflows on annotation platforms.
• Utilize SQL, Python, or dashboards to monitor pipeline telemetry, analyze rater quality, and diagnose data anomalies.
• Navigate ambiguity to build consensus across research, engineering, and legal, managing multimodal and internationalized data workflows.

• Bachelor’s degree in a technical, quantitative, or operational field, or equivalent practical experience.
• 5 years of experience in technical program management, data operations, or AI/ML program management.