Mark A. Smith.
AI, data-science & analytics product manager — turning models, pipelines and agents into things customers actually feel.
AI, data-science & analytics products,
forged in a regulated industry.
20+ years shipping data and AI products — currently leading an AI product portfolio (generative, agentic, predictive) inside an investor-owned utility. Earlier: stood up the enterprise data-science platform and governance program at EPRI, built Reliant's customer-facing analytics suite (disaggregation, anomaly detection, premise-level forecasting) on a Hadoop / Spark big-data platform, and ran ML pipelines at HP. Engineering foundation in systems and industrial engineering; the electric sector has been my proving ground, not the limit of the playbook.
A single line, drawn across four utilities and three decades.
- 2023–NowEntergyAI Product ManagerGenAI · Agentic · Predictive
- 2021–22Gexa / NextEraDER Product / AnalyticsSegmentation · Optimization
- 2017–20EPRIData Science LeadPlatform · Governance · MLOps
- 2002–16Reliant / NRGPrincipal, Product InnovationBig-data · Customer analytics
- 1990s–17HP · Questia · CompaqEarlier CareerML pipelines · DQ · Six Sigma
Two artifacts, both about how AI gets built.
One is a three-book series on enterprise AI delivery. The other is this site — built on Lovable as a live demo of the methodology those books describe. A gallery of additional Lovable apps lives on the projects page.
Enterprise Agentic AI
Three books on AI-assisted software delivery, Specification-Driven Development, and citizen-AI governance. Published at microwaterfall.com.
This site, built on Lovable
A working demo of the Micro-Waterfall methodology. Specs → prompts → routes → ship, on a single afternoon.