Building production AI systems. Teaching the engineering behind them.
I work across agentic AI, conversational analytics, semantic layers, machine learning platforms, and cloud architecture. This site is where I publish the systems I build and the classes I’m developing around them.
Conversational Analytics with Looker
A practical course for engineers, analysts, and architects who want to move beyond “chat with data” demos and understand the semantic, evaluation, and production layers underneath trustworthy conversational analytics.
From semantic layer to trustworthy data agent
LookML foundations, Explores, data agents, instructions, verified queries, evaluation, custom APIs, auditability, and enterprise deployment.
Built to expand beyond one class
Production Agent Engineering, Mathematical Foundations for ML, and Production ML on GCP are planned as dedicated tracks.
What I’ve built
A few systems that represent the work I do now and the platform engineering behind it.
Enterprise Agent Orchestrator
ADK, A2A, MCP, capability discovery, routing, policy, evaluation, and enterprise controls.
AnalyticsConversational Analytics Platform
Looker Core, LookML semantics, data agents, instructions, evaluation, and custom enterprise integration.
ML PlatformMLevis on Vertex AI
Modernized ML orchestration from hosted Airflow DAGs to Vertex AI pipelines.
Feature PlatformFeature Store Modernization
Migration toward Vertex AI Feature Store 2.0 with simpler operations and lower cost.
Product MLHaloAI
Virtual assistants, messaging automation, discount decisioning, intent routing, IVR/ASR, and real-time inference.
CloudMulti-cloud Modernization
Consolidated AI/ML, analytics, and data engineering workloads from multiple platforms toward GCP.