Category: Data
-
Antifragile AI Architectures
AI is full of contradictions: capable but unreliable, local improvements create externalities, generalist models are evaluated against specific criteria, and so on. Antifragility is a framework that deals in contradictions too, and seems an appropriate lens through which to explore AI systems architecture, as I had used it in an earlier era to explore hand-crafted…
-
Data and AI mini blogs
A brief reflection on Thoughtworks Australia Data & AI mini-blogs. From May 2021, we set ourselves a target to publish a short blog on a different AI, ML or data topic every week. Here’s the pitch from the landing page: Bite-sized content delivering valuable insights from Thoughtworkers who wrestle varied client problems week in and…
-
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…
-
trippler multi-destination
It’s time for my EV trip planning app trippler to cater to multiple destinations, beyond the beaten track of simple A-to-B trips. This is another feature I decided I needed when planning and driving The Cross to The Cape. Multiple changes I’ve wanted multiple destinations for some time, but recent refactoring of the UI and…
-
trippler contingencies
Planning a dawn-to-dusk adventure from snow to surf in an electric vehicle meant not just planning for resilience, to allow for changes to the plan, but also planning for contingencies, to know in advance exactly how to respond to changes. Carnival of carving We lay our scene on the descent from Mount Hotham to Cape…
-
trippler at PyConAU
I was thrilled to be back for my second PyCon AU – with a wonderfully diverse and inclusive group of technologists – presenting on trippler in a talk titled An EV Trip Planner for Australia. I got a great feedback, including a suggestion to incorporate the many very Australian BIG things we might encounter on…
-
trippler for Aotearoa New Zealand
Planning to present my PyCon AU talk to an internal audience at MYOB, I realised the title An EV trip planner for Australia, while entirely appropriate for an Australian conference on Python, wasn’t as inclusive as it could be for the members of a technology organisation encompassing Australia and Aotearoa New Zealand. So trippler now…
-
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…
-
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…