I have spent my career building the systems beneath intelligent products: how data becomes usable, how models reach production, and how developers build on the resulting platform.
Unstructured — data infrastructure for AI products
I lead product across a portfolio that helps teams turn unstructured information into reliable AI applications. The work spans the core data pipeline, agent-accessible knowledge, and developer experiences.
Pipelines
Productizing document-processing infrastructure for production AI workflows: APIs, platform experiences, deployment paths, and enterprise use cases.
Foundation
AI infrastructure for enterprise agent search and context. Foundation gives AI agents access to relevant enterprise knowledge through a single MCP-native interface—improving retrieval without treating ever-larger context windows as the solution.
Transform
Reducing the distance between a developer’s first file and a working AI-data workflow.
Amazon — production ML platforms
I led product for an internal ML deployment platform that connected applied science with production software. The work spanned model lifecycle, experimentation, orchestration, and the systems needed to make ML reliable for real users.
Amazon and Expedia — enterprise data platforms
I led data-platform work in large engineering organizations, helping teams move from fragmented or legacy data systems toward reliable, scalable foundations for analytics and software development.
Open source and developer tools
I stay close to implementation through open-source contributions and practical tools. Recent work includes MCP integrations and a local, read-only MCP for personal-finance data—designed to make AI assistance useful without granting unnecessary write access.
Sudarshan Sampath 