I build the pipelines that make AI products actually work in production.
Senior Technical Program Manager repositioning toward AI Product Management — hands-on with RAG pipelines, MCP integrations, and agent tooling that ships, not just prototypes.
From technical program management into hands-on AI product work.
I'm a Senior Technical Program Manager at Kantar, a B2B2C survey platform serving enterprise clients, where I lead AI product initiatives end to end — from OGSP planning and multi-region launch coordination to the underlying retrieval and agent infrastructure that makes those products work.
Over the past year that's meant building a shared RAG knowledge base on pgvector, wiring agents into real systems of record through MCP, and shipping an autonomous investigation agent — Jira-Bug-Investigation — that cut average triage response time from 48 hours to under 1 hour, and fix-to-PR cycle time from a week to 2 hours.
- Agents Claude, MCP, Claude Code, Codex
- Retrieval pgvector, OpenAI embeddings, New Relic, S3
- Systems Atlassian, GitHub, Figma, Lucid MCPs
- Program OGSP planning, RICE prioritization, multi-region launch
- Analytics Python, SQL, Tableau, Power BI
Agent and retrieval systems, shipped.
Graph-Engineer: DAGs of Sub-Agents
→A Claude Code skill that turns a prompt into a validated graph of parallel sub-agents instead of one long serial run
Jira Triage Agent
→An LLM agent that reads tickets, searches the codebase, and cross-references logs before an engineer ever opens the ticket
RAG Knowledge Base
→A pgvector-backed retrieval layer over Confluence and S3, built to give internal agents an actual memory of how the product works
MCP Tool Integrations
→Wiring agents into real systems of record through the Model Context Protocol