Category: Machine Learning
-
LLMs are lineage black holes
Data lineage is important to most organisations, even if they don’t make use of it. Systematically capturing the upstream provenance and downstream consumers of any piece of data is critical to trusting the utility of that data and understanding its impacts, at any scale beyond a handful of excel spreadsheets. The nature of lineage When…
-
LLMs as text simulators
I’ve often written here about developing systems that leverage simulation. Simulation combining physical processes, information systems, and crowd behaviours. Simulations that support organisational decision making, customer experiences, and learning. And in late 2025 we were having a moment where more people were starting to describe Large Language Models (LLMs) as simulators of text. Simulation of…
-
GenAI in Data Platforms
I was part of a panel on the Impact of GenAI on Modern Data Platforms recently, hosted by the Data Engineering Melbourne meetup. It was great to chat with MC Ryan Collingwood and fellow panellists Rahul Trikha, Peter Barnes and Tony Nicol in front of a large and curious crowd. Like the crowd, I felt…
-
Effective ML Teams in Korean
We received our copies of the Korean translation of Effective Machine Learning Teams this week!
-
Team Topologies x EMLT
To provide more resources around the excellent conversation we had with Matthew Skelton on the Stream of Teams podcast in late 2024, we collaborated on an article summarising our chat, which was based on the final chapter of Effective Machine Learning Teams (EMLT). The article titled Team Topologies in action: Effective structures for Machine Learning…
-
22 rules of generative AI, 2 years on, Ghibli intermission
This update comes at the peak of the Ghiblification fad. There’s nothing new here and yet I found the release of and response to GPT 4o image generation, including mechanistically crushed and artificially reconstituted Ghibli, particularly shocking. So here’s a retrospective case study for the recent review of rule #10 (labelling ingredients) and a longer…
-
22 rules of generative AI, 2 years on, part 1
How has my original post on 22 rules of generative AI aged in a period of rapid change? Are these solution considerations as enduring as I thought? Let’s reflect on the original advice and developments in the meantime. Apologies again for minimal references as I timeboxed the writing. In general, evidence should be discoverable/verifiable with…
-
AI conversations for every role
We always envisaged Effective Machine Learning Teams (EMLT) speaking to multiple roles. Recently we’ve spoken to 4 distinct audiences about ML development, and you can listen in too. Whether you identify with: there’s a podcast* for you with popular presenters – see more below. Product We joined a Product Tank Melbourne event, along with Ana…
-
Joe Reis show
I had a great chat with Joe Reis on The Joe Reis Show, along with one co-author of Effective Machine Learning Teams, David Tan. We covered the motivation for the book, scenarios where we’d helped teams address issues to become more effective, and the tools we’d developed from those solutions. We also talked about a…
-
Effective topologies for Data and ML teams
I presented this talk at the Melbourne Data Engineering meetup, on a wild and wet Friday night. Having your cake and eating it too – Effective topologies for Data and ML teams (slides) In the talk I explore how Team Topologies provides patterns for reconciling fast flow of value with (multiple) specialisations in data and…