AI Data Strategist

Dyna Robotics builds general-purpose robots powered by a proprietary embodied AI foundation model with top-in-industry generalization and real-world performance. Already deployed with customers across multiple industries, our robots do commercial-grade work in the physical world. Our team comes from Google DeepMind, Meta, and Cruise, and we're backed by CRV, First Round, and other leading investors.

The Role

We are hiring an AI Data Strategist to define the data requirements that drive model improvement across Dyna's robotics platform.

This is a senior individual contributor role that focuses on strategy rather than managing operational execution. Instead of running the day-to-day data pipeline, you will define what operations and research execute against. You will establish the specifications, frameworks, and feedback loops that determine whether our data actually improves our models.

The core question you will help answer every week is: our model failed here, so what does that mean for our data strategy?

What You'll Do

1. Define Data Collection Priorities

   • Identify lifecycle gaps: Maintain a clear, comprehensive view of where the data lifecycle has gaps, from pre-training through post-training.

   • Direct collection efforts: Prioritize what the data collection team should focus on next, clearly distinguishing between data that merely adds volume and data that actually drives model performance.

2. Design Evaluation & Quality Frameworks

   • Set the standard: Define how robot episodes should be labeled and determine what rubrics and taxonomies capture meaningful signal.

   • Establish quality benchmarks: Define what "good data" looks like for each task and model stage so the labeling team can execute flawlessly against your standards.

3. Extract Signal from Operations

   • Translate field realities: Partner closely with the operations team to understand what is happening in the field, including shift handoffs, collection quality, and deployment issues.

   • Inform data strategy: Act as a strategic consumer of operations output, translating real-world operational realities into high-impact data strategy decisions without directly managing the operations team.

4. Build Data Lifecycle Observability

   • Define health metrics: Establish the metrics that measure the health of each phase of the data pipeline, including collection coverage, label quality, evaluation consistency, and model feedback loops.

   • Drive visibility: Create a real-time, organization-wide view of data lifecycle health.

Who You Are

• Systems Thinker: You understand that superior models come from exceptional data strategy, not just massive data volume.

• Structured Problem Solver: Highly analytical and detail-oriented, with the ability to translate messy, real-world failures into structured frameworks.

• Analytically Minded: Possess strong instincts for failure analysis, dataset structure, and the feedback loops between deployment and training.

• Cross-Functional Influencer: Able to rally and influence cross-functional teams without needing direct authority.

• Clear Communicator: Strong written and verbal communication skills, with the ability to prioritize effectively in fast-moving environments where everything feels urgent.

What You’ll Bring

• Core Experience: 4-8+ years of experience working in AI/ML, robotics, autonomy, or data-centric systems roles.

• Data Strategy Expertise: Proven experience defining data quality standards, evaluation frameworks, annotation systems, or data strategy for machine learning products.

• Collaborative Track Record: Experience working closely with cross-functional teams, including ML researchers, operations, annotation teams, and engineering.

• Edge-Case Proficiency: A deep understanding of how deployment failures, edge cases, and real-world operational data translate into model training and evaluation improvements.

Bonus points for

• Experience operating in fast-moving, ambiguous startup or R&D-heavy environments

• Experience with embodied AI, video, or time-series data.

• Familiarity with evaluation pipelines, active learning, or data-centric AI.

• Exposure to annotation tooling such as Labelbox, Scale, CVAT, Encord, or Voxel51.