David Colls

Occasional tweeter and a ThoughtWorker. All views shared are my own.

[Next Day Video @nextdayvideo]

#pyconau @davidcolls Maths Whimsy with Python https://t.co/ZMECxfWLug

[Thoughtworks @thoughtworks]

Integrating privacy into your data science workflow is still complicated. In her forthcoming book @kjam will give you solid advice and best practices on breakthrough privacy-enhancing technologies. It's currently in early release: https://t.co/sOpMRSExb3

[Thoughtworks @thoughtworks]

On this joint mission with @REA_Group, we helped develop Australia's most accurate property valuation tool by bringing our global experience in #data and #AI to complement the client's intimate domain knowledge of applying #MachineLearning in property valuations.

[Thoughtworks @thoughtworks]

Discover how to define and deliver greater value in your next data engineering project in our new Modern #DataEngineering playbook - check it out on our website: https://t.co/IU4eMH4bji

[Thoughtworks @thoughtworks]

We’re ready to geek out with you at @dataengconfau in Sydney today 👋 pop by our booth and grab a copy of our Modern Data Engineering Playbook - see you soon! #TechConference #Data

[David Colls @davidcolls]

I had fun making a second Semantle solver. Not ready for live gameplay but good for offline strategising. @SemantleGame https://t.co/zL0StPggtI

[QUAD LOCK @QuadLockCase]

On June 26th, Melburn Roobaix - The Hell of the North(cote), was back for 2022! After 2 long years of hiatus due to lockdowns in Melbourne, we were excited to be a part of the day. @fyxo. 👏🏻
#melbourne #cycling #roobaix #event

Read Now: https://t.co/PVnd8c3lri

[David Colls @davidcolls]

I made a Semantle solver. https://t.co/ApZFDNgaZV

[Andy Matthews @stupidoldandy]

There are three days remaining to chip in to the @stupidold moving fund, if that’s the kind of thing you’re into. A colossal thank you to the 626 deadset legends who have helped out so-far. I am in awe.

https://t.co/5EIW1m4i8Y

[Jez Humble @jezhumble]

Since this false assumption is still distressingly common: In software delivery, there is no speed vs stability trade-off

We now have 7yrs of data from tens of thousands of responses globally. Attached graph from https://t.co/DeiFnyjuNZ, previous years at https://t.co/NKa8umZOo1

[DataEngBytes Conferences @dataengbytes]

Up next, we have @nletcher taking us through Metrics-driven Data Architectures:
#DataEngBytes #dataengineering #metrics
https://t.co/rIzuo1nYoo

[David Colls @davidcolls]

The slackometer - part 2 of my lockdown slackline pose estimation project https://t.co/lLP2f3ryeD

[Ned Letcher @nletcher]

So metrics layers huh?

Come to my talk at #DataEngBytes and hear what all the buzz is about. Is it a shiny new piece of the modern #data stack? Or is it OLAP cubes 3.0? https://t.co/NE7gnWPGB4

(spoiler: it's probably both)

[Simon Aubury @SimonAubury]

Fantastic #dataengineering speaker lineup at
@dataengconfau next week. #DataEngBytes! is a 3 day community conference centred on real world data engineering problems and solutions - with an 🇦🇺 friendly timezone.

🎟️ Amazingly - this stuff is free! https://t.co/IbzEjBEiu5

[sPam Rucinque @PamRucinque]

Shout out to the incredible @ineffyble for creating this masterpiece.

[David Colls @davidcolls]

Spent far too long on a mechanical odometer display in @matplotlib this weekend, but fairly pleased with the result https://t.co/KTbnLTHEq3

[Sarah Taraporewalla @sarahtarap]

Thank you to @ben_caunt and the rest of the crew for a wonderful week of #obliquefest. Just one of the reasons I ❤️ @thoughtworks so much - people like you going out of your way to fill our souls and still our minds 😃

[David Colls @davidcolls]

Another lockdown, another project for body and mind. https://t.co/bFKoHjH72i

[Harmeet (She/Her) @hksokhi]

Data engineering Melbourne Meetup today at 6:00 pm AEST : David Colls would be sharing Architecture of Data Mesh , it's building blocks and some lessons we have learned . Join us at https://t.co/OjIZl3rpao https://t.co/uCyEGfiNP8

[David Colls @davidcolls]

@freddy_fostvedt In general, you can trade off raw “opaque box” performance for interpretability of models. You can reflect this to the consumer to build trust. There is a lot of research currently in this area. This post gives an overview of some techniques
https://t.co/axMAqlmBuB 6/6

[David Colls @davidcolls]

@freddy_fostvedt In model design, we talk about interpretability and explainability. This is something I probably should have added to the achievements unlocked yesterday, with appropriate discussion - AI will learn faster, more responsibly, and _with more explanations_! 5/6