Senior AI Data Scientist, Agentic Automation (Marketing)

As Senior Data Scientist in the Applied AI team, you will build agentic marketing-automation systems that span multiple brands and languages. You map processes, quantify ROI, and deliver production-ready solutions that operate end-to-end and credibly justify decisions. You’ll shape auditable, self-documenting workflows and deploy them with guardrails, ensuring safe, scalable influence on marketing outcomes. This role blends data science with software engineering to drive measurable impact across SMB brands.

• Map current marketing workflows and attach hours per step to quantify effort
• Extract live requirements from non-technical stakeholders
• Design state transitions, API calls, and failure handling across steps
• Build guardrails: dry-run mode, approval gates, least-privilege scopes, undo mechanisms
• Define human-in-the-loop policy and review queues for marketers
• Write evaluation designs for non-traditional metrics (precision/recall, source verification, brand voice checks)
• Validate tracking data quality and quantify errors before production
• Integrate agents with webhooks or scheduled triggers and ensure idempotent behavior
• Decide when to use frontier vs cheaper models and prove cost/latency tradeoffs
• Evaluate vendor APIs, rate limits, data models, integration costs
• Ensure safety: staging modes, approvals, auditability, and rollback paths
• Ship end-to-end: code, containerise, instrument, deploy with minimal DevOps support

• 7+ years building data and ML systems
• Expert in Python and ML
• End-to-end production experience with multi-step, tool-calling LLM workflows
• Experience integrating with third-party APIs, including auth, rate limits, pagination
• Cost and latency optimization in model routing and caching
• Safety-focused design with staging, approvals, least-privilege scoping
• Evaluation design for generative/agentic output (calibration, regression suites, red-teaming)
• Process mapping and quantitative analysis of workflows
• Technical vendor evaluation (APIs, data models, integration cost)

• Clear communicator with domain experts
• Strong problem-framing and requirement extraction
• Judgment and accountability for production decisions
• Python
• Machine Learning
• Docker/Containerization