Head of Supply Chain

About The Role

We’re building AI agents that give supply chain operators real-time intelligence about data quality, risk, and reliability - before bad data causes a costly downstream decision. These aren’t dashboards or BI wrappers. They’re agents that understand BOM hierarchies, supplier data feeds, and demand signals deeply enough to reason about failure modes, score data risk, and generate validation and remediation playbooks automatically.

We’re building AI agents that give supply chain operators real-time intelligence about data quality, risk, and reliability - before bad data causes a costly downstream decision. These aren’t dashboards or BI wrappers. They’re agents that understand BOM hierarchies, supplier data feeds, and demand signals deeply enough to reason about failure modes, score data risk, and generate validation and remediation playbooks automatically.

Every large manufacturer depends on data from contract manufacturers, logistics partners, and internal systems - purchase orders, inventory records, demand forecasts, BOM structures. That data is almost always incomplete, inconsistently formatted, or wrong in ways invisible until someone makes a multi-million dollar decision on top of it.
The traditional answer is a validation team running SQL checks. The AI answer is an agent that reads the data specification, understands the business process it supports, identifies every field-level failure mode, scores severity and risk, and writes the validation rule set and remediation playbook - automatically, at a depth and speed no analyst can match.

• Design and spec AI agents for supply chain data intelligence: FMEA automation, validation rule generation, data quality scoring, anomaly detection, and remediation workflow design.
• Translate raw manufacturer and supplier data specifications, BOM governance documents, and data dictionaries into structured LLM context, few-shot examples, and domain grounding.
• Own the data governance framework: define what good data looks like across supply chain domains and build the severity/occurrence/detectability rubrics agents use to prioritize problems
• Build working LLM prototypes:
  Use Claude/Lovable or similar tools to build working prototypes to gather feedback and communicate requirements.
• Run customer discovery workshops onsite: map data landscapes, identify highest-risk data elements, return with a prioritized validation roadmap and working proof-of-concept.
• Build the enterprise deployment playbook: COE model, wave rollout approach, training materials, and executive narrative for scaling across business units and geographies.

• Not a data engineering role - you are building the intelligence layer, not the infrastructure beneath it.
• Not a pure PM role - you will build prototypes
• if you haven’t yet spent time building with LLMs, this will be a stretch.
• Not a consulting engagement - you are building a product and a company, not delivering a project and moving on.
• Not for someone who needs a team beneath them - you are the first hire
• the second hire reports to you.
• Not a role where domain knowledge substitutes for technical ability, or vice versa - both are required.

Days 1-30:
Shadow 2-3 customer supply chain teams; document data feeds, failure modes, and manual workarounds; produce a prioritized problem inventory.
Days 31-60:
Ship a working AI agent that processes a real customer data spec, scores risk using FMEA logic, and outputs a structured validation playbook.
Days 61-90:
Write the product spec for v1, define the COE rollout model, identify the first three expansion accounts, present the go-to-market roadmap to the founding team.

Supply chain data governance has…