Talks, workshops, advisory
Speaking & advisory
I speak about the organizational half of data as much as the technical half, because that is where the hard parts have consistently been. Every topic below has a written case study behind it, so you can read the long version before deciding.
Anil Thapa
Data & AI Platform Leader
Data platforms and the teams that run them. Functions built from zero, warehouses consolidated out of silos and mergers, and a lot of time spent convincing executives that a governed number is worth what it costs to produce.
Topics
Building a data function from zero
Centralizing scattered analysts without becoming the bottleneck they were routing around. Definitions first, people second, tools last, close to the reverse of how it is usually run.
Read the long version 02What changes when data reaches the boardroom
Executives do not want the number, they want the decision it implies. Presenting uncertainty without hedging, disagreeing without losing the seat, and answering the question behind the question.
Read the long version 03Why nobody believes your numbers
Trust is a step function, not a gradient. The four failure modes that actually break it, why coverage is the wrong goal, and how to run an incident so it costs you less than the error did.
Read the long version 04Agentic systems against a mature warehouse
Frontier models score around 86% on clean benchmark schemas and closer to 10% on real enterprise ones. What breaks in that gap, and which guardrails hold up in production.
Read the long version 05What your data platform costs, and who decides
Attribution before optimization, the structural causes of waste that recur everywhere, and which inefficiencies are worth leaving alone on purpose.
Read the long version 06Merging warehouses, and the definitions underneath
Mergers, departmental silos, platform migrations. Why definition reconciliation is the project and the pipeline work is support: a political problem wearing a technical costume.
Read the long versionWays to work together
Conference talks & panels
Any of the topics above, sized to your programme. I will tell you quickly whether it fits rather than leaving you waiting.
Podcasts & interviews
Long-form suits this material better than a stage does: the tradeoffs are the interesting part and they need room.
Team workshops
Working sessions with a data team on their actual problem: definitions, trust, platform cost, or how to structure the function.
Advisory
A second opinion before something expensive gets committed
The pattern I see most: good people, a real budget, a platform that works, and nobody happy with what comes out of it. That is rarely a technology problem. It is usually that nobody has agreed what the numbers mean, or that the function is shaped wrong for the company it serves, and both are expensive to discover late.
In the problem
A specific decision with a deadline on it. Reviewing a platform or warehouse consolidation plan before it is committed, pressure-testing an architecture, or working through what a migration will actually cost to run.
Alongside the leader
Standing sounding board for a data leader. Structuring a new function, working out what to hire and in what order, how to hold a definition when a stakeholder wants it changed, and what to do about the platform bill.
Worth saying plainly: I do this alongside a full-time role, so I take on very little of it. If you want to know how I think before you ask, the case studies are the long version and every one of them states what the decision cost.
For organizers and teams
Let's talk about what you need
Tell me the date, the format and who is in the room, and I will come back with a yes or a no quickly rather than leaving you waiting. Short bio, headshot and abstracts available the same day.