Xander Bailey

Software Engineer. Medium: he/him/his

[Akshay Krishnaswamy @hyperindexed]

Exquisite engineering, across this work...

In addition to contributing to the open Iceberg project, the team has preserved interoperability as they have extended core capabilities.

e.g., branching that isolates schemas, partition specs, and sort orders: https://t.co/twaAAiZBn2

[Xander Bailey @xanderrbailey]

@andrewlamb1111 @ApacheDataFusio https://t.co/M76evrLqQO

[Xander Bailey @xanderrbailey]

@andrewlamb1111 @ApacheDataFusio We’re excited to be launching this to all users very soon!

[Xander Bailey @xanderrbailey]

@kimmonismus Don’t get me wrong, it what OpenAI is promising is true then this is very powerful but AGI, it is not. It’s still spicy autocomplete.

[Luke Bailey @LukeBailey181]

Can interpretability help defend LLMs? We find we can reshape activations while preserving a model’s behavior. This lets us attack latent-space defenses, from SAEs and probes to Circuit Breakers. We can attack so precisely that we make a harmfulness probe output this QR code. 🧵

[Xander Bailey @xanderrbailey]

@JsonBasedman Hi @JsonBasedman, I’m from the Pipeline Builder team and I’m pleased to let you know that we’re working on the feature and it’s nearly ready for you to try! If you post on the community page, we can keep you up to date there! https://t.co/y1ABRYf1Or

[Xander Bailey @xanderrbailey]

Since we started working on productising generative AI workflows we’ve been eagerly waiting for an open-source model that’s truly competitive. Llama 3.1, sounds like it could be an inflection point when it comes to cost and performance. https://t.co/UM0qujHpc7

[Palantir @PalantirTech]

Go beyond chat. Maximize production.

AIP in action.

[Xander Bailey @xanderrbailey]

Fighting with #GPT to tell me the built in system message.

[Xander Bailey @xanderrbailey]

@raymondh @dataclass “Explicit is better than implicit”, this only saves me a few lines of code and it’s harder to read no?

[Kareem Carr, Ph.D. @kareem_carr]

this really hurts

[Sebastian Raschka @rasbt]

Well, this was unexpected. I guess I am not a double-descent skeptic anymore then ... https://t.co/awGZjC3Mcp

[Xander Bailey @xanderrbailey]

Gaussian Process to predict total vaccination numbers in the UK (first dose). Key number is 15 million which represents the top 4 most vulnerable groups. We’re due to hit that between 07/02 and 14/02! #bayes #MachineLearning #COVID19 #vaccine #PfizerVaccine #astrazenica

[Xander Bailey @xanderrbailey]

New article published in @TDataScience about modelling tennis game outcomes using Bayesian modelling. Let me know what you think!
#MachineLearning #mcmc #statistics #DataScience https://t.co/fYWxXIQjeQ

[Towards Data Science @TDataScience]

Variational Inference for Neural Networks by @xanderrbailey https://t.co/GxerK6rbAg

[Xander Bailey @xanderrbailey]

New article published in ⁦@TDataScience⁩ “Variational Inference for Neural Networks” #MachineLearning #deeplearning #statistics #python #pytorch
https://t.co/12XCrm7886

[Xander Bailey @xanderrbailey]

I just published in @TDataScience: Introduction to MCMC. What is the Metropolis algorithm and why does it work ?https://t.co/fJxeXzPs97 #machinelearning #ml #datascience #python

[Xander Bailey @xanderrbailey]

NHS contact tracing app is finally available! Check out my blog post about how the technology works and how it keeps your data safe!

#covid #nhs #contacttracingapp
https://t.co/KJ0NYYEH5R

https://t.co/RStVhmpfFG

[Barack Obama @BarackObama]

Justice Ruth Bader Ginsburg fought to the end, through her cancer, with unwavering faith in our democracy and its ideals. That’s how we remember her. But she also left instructions for how she wanted her legacy to be honored. My statement: https://t.co/Wa6YcT5gDi

[Xander Bailey @xanderrbailey]

Nice thing I learned today when building a @pythonflask app.

Do the authentication in "app.before_request" rather than each end-point.

This shuts down access to any non-implemented endpoints that might still be accessible from the web #api #python #flask