Speaking
Speaking
I speak on analytics agents, the semantic layer, data foundations for enterprise AI and the finance data modeling underneath all of it. Practitioner-level. First-person. No slides full of buzzwords.
Topics
The semantic layer is the API for AI
Why text-to-SQL demos well and fails quietly on finance data, and how approved metric definitions, declared grain and a single tool over MCP turn an analytics agent into something a CFO can rely on.
Read the essayEvals for Analytics Agents
Making an agent’s accuracy a number with a date on it: golden question sets, ground truth from certified models, deterministic scoring, refusal recall and regression gates in CI.
Read the essayData Foundations for Enterprise AI
Working backwards from the agent to the pipeline: ingestion patterns per source, three-layer modeling, and the data-quality checks at every layer that keep a green pipeline from producing wrong numbers.
Read the essayFinance Data, Done Properly
Ledger-first architecture, bitemporal reporting, basis-point fees on moving balances and chart-of-accounts migrations. The modeling decisions that make the numbers agree with finance.
Read the essayAgents in the Loop, Not in Charge
Production patterns for coding agents inside data teams: deterministic gates first, sandboxed tests, traceability, and an agent that triages but never applies the label.
Read the essayEvaluating Frontier and Open-Weight Models for Real Workflows
How to choose a model from the business problem outward: the quality the task demands, where data is processed, what it costs per successful task, and who operates the result.
Read the essay
Invite Me to Speak
I speak at conferences, company offsites, podcasts and private workshops. If you'd like me to present on any of the topics above, or something adjacent, get in touch and tell me about the audience.

Past Appearances




