You already run Purview. DLP policies, sensitivity labels, auto-labeling, SITs, Endpoint DLP, maybe DSPM for AI. Then someone pastes a customer record into Claude, Copilot desktop, or ChatGPT, and none of that policy is fired.

That is the job.

Wald is the on-device AI DLP layer for that gap. We inspect prompts and file uploads locally, classify with context instead of regex, and redact before data leaves the laptop. We do not replace Purview. We extend the program you already built into AI tools Purview was never designed to see.

We are a small, SOC 2 Type II team in Palo Alto and India, used by 55+ regulated organizations. You will be the person who can sit with an M365 / security / compliance owner, speak Purview fluently, and show where Wald maps onto their labels, DLP, and Copilot controls.

What you will do

- Own technical delivery in Microsoft E5 / Purview environments: discovery, architecture, POC, and rollout
- Map existing sensitivity labels, SITs, trainable classifiers, and DLP policies onto Wald so customers keep one classification language
- Run working sessions with security, compliance, legal, and M365 admins — not slideware




- Design reference architectures for Copilot, ChatGPT, Claude, Gemini, and shadow AI on the endpoint
- Feed product what is actually deployed in those tenants: label taxonomies, false-positive pain, DSPM for AI limits, Endpoint DLP gaps
- Write the artifacts Purview people trust: policy mapping docs, KQL-style investigations, rollout plans, admin guides
- Be a peer to practitioners. The first conversation with a candidate is often a working discussion of their current Purview stack

This role is a fit if you have
- Hands-on Purview, not just a cert: DLP (Exchange, SharePoint, OneDrive, Teams, Endpoint), Information Protection / sensitivity labels, auto-labeling, SITs or EDM, retention or records
- Shipped this in a real tenant — enterprise, or consulting across several
- Enough M365 to be dangerous: Entra ID, Exchange Online, SharePoint, Intune or Defender, Graph or PowerShell
- Comfort in front of CISOs and M365 admins without a script
- Curiosity about Copilot, DSPM for AI, and why prompt-level DLP is a different problem from file/email DLP

📌 Senior Solutions Architect, Microsoft Purview & AI Data Protection (India)
🏢 Wald
📍 India