Speaking
Talks on AI, ad tech, data science and startups
Keynotes, panels, workshops and internal sessions. The through-line is the same as the rest of this site: what the technology actually does when you put it into a real business, rather than what the deck says it will do.
AI and transformation
- What AI changes about running a business Where AI is already paying for itself, where it is still a science project, and what has to change in how a company is organised before either becomes true. Includes the uncomfortable part: which functions get smaller.
- What AI changes about building software Agentic coding tools have moved from demo to daily use, and the effects land on team structure, code review, hiring plans and what a senior engineer is for. A working account rather than a forecast.
- Agents in practice What an agentic system does step by step, where they break, what they cost to run, and how to tell a real deployment from a wrapper around a prompt.
- Buying AI without being sold to How to evaluate vendor claims, what benchmarks are worth, and the questions that separate a working product from a convincing pitch.
Ad tech
- Custom bidding and algorithmic buying Bidding against your own business KPIs rather than the platform's proxies — what it takes to build, what it moves, and why most advertisers leave it on the table.
- Supply path optimisation Which paths to a piece of inventory are worth taking, what the extra hops cost, and how to run SPO as an ongoing discipline instead of a one-off audit.
- Curation What curated marketplaces actually change about price and quality, who captures the value, and how to tell a genuine curation product from a repackaged reseller.
- Fraud and made-for-advertising inventory What MFA and fraud cost a campaign, why the incentives keep both alive, and what a buyer can do about it without pretending the problem is solved.
- Agentic media buying Where AI agents are moving into planning and activation, what they automate well, and what still needs a trader.
- Measurement, incrementality and carbon Whether the numbers a campaign reports mean anything, how to measure lift honestly, and how much carbon the supply chain emits along the way.
Data science
- Building and scaling a data science function What the first hires should be, when to specialise, how the function earns its keep, and the failure modes that show up at every size.
- Why models underperform The usual suspects — target leakage, drift, the wrong objective, a business problem nobody translated properly — and how to diagnose which one you have.
- Hiring data scientists Writing a role that describes the actual job, sourcing beyond the obvious pools, and assessing candidates without a whiteboard theatre.
- Behavioral economics for decision-making What twenty years of decision research says about how people actually respond to information, and what that means for both product design and how you present analysis internally.
Startups
- Getting a company off the ground The whole arc, from an idea worth testing to a business that can pay people — what to build first, what to leave undone, and which early decisions are expensive to reverse.
- The founding team Who you actually need at the start, how to divide the work between people who all think they're the technical one, and what the first hires outside the founders should be.
- Raising money as a technical founder What investors ask about the technology, what they mean by it, and how to answer without either overclaiming or disappearing into detail.
- Selling into ad tech as a small company Finding the first customers in an industry that buys on relationships, getting through procurement, and surviving the pilot that never ends.
- Exits and being on the other side of diligence What acquirers examine, how technical diligence actually runs, and what to have in order long before anyone asks for it.
Booking
Conferences, industry events, company all-hands, board sessions and engineering-team workshops. Talks are rebuilt for the audience rather than delivered off the shelf, so tell me who is in the room and what you want them to do differently afterwards.