I'm a martech and CDP architect who spends most of my time in the gap between marketing intent and the data systems that have to make it real — identity resolution, event pipelines, and lately the AI agents that act on top of them. My path runs through LG, Samsung, and a run of earlier ecommerce and performance-marketing roles, always ending up as the person who builds the plumbing everyone else assumed already existed.
- Reusable over one-off Data infrastructure gets built once and reused for the next onboarding, not rebuilt per request — that's what made a 4-country CDP rollout possible in 1.5 years with a 3-person team.
- Governance before scale Identity resolution and golden-record rules get defined before a platform goes live, not patched in after the data gets messy.
- Automate the boring parts If a workflow repeats, it becomes an n8n agent or a scheduled pipeline, not a recurring line item on someone's task list.