Vaishnavi Hardikar

Enterprise account · Turnaround

Rescued a failing enterprise account

Brought in as incident and problem reports were spiraling. I rebuilt the API design framework from the ground up: reference "golden" designs as a quality standard, Spectral linting rules to enforce it automatically, a full test suite, documented exceptions, and a code-generation-aligned template.

I was both the technical and the functional expert on the account, owning the fix end to end.

AI platform · Production

Built the AI layer of an API platform

Shipped two production capabilities: AI for API (LLM-assisted API design with automated Spectral linting) and API for AI, an agent-to-backend layer defining how agents access data, which tools they can call, and the safety constraints around them. Added a RAG retrieval layer over consolidated user guides for a customer-facing agent.

Built a custom MCP server connecting AI agents to Confluence for governance decision records, and helped set the standards for how MCP servers are built and consumed across the division. I've since rebuilt that pattern from scratch in public: github.com/vhardikar/confluence-mcp ↗

Data governance · EU AI Act

Led AI data governance to EU AI Act standards

As R&D Data Officer for the division, I owned data lifecycle governance aligned with EU AI Act Article 10. I designed the end-to-end pipeline behind an AI Advisor: roughly 25 providers consolidated through MDM deduplication into golden records, with human-expert evaluation and an API gateway.

Enablement · Zero to one

Stood up an enablement function from zero

Built the integration and advocacy function 0→1: integration guides, training catalogs, self-serve diagnostics, RACI governance, and a network of local governors I recruited and trained. In parallel, I ran the end-to-end onboarding of six-plus acquired business units, replacing 10+ bespoke integrations with standardized, repeatable templates along the way.