aburla

4 ms·

1.
by aburla 4y ago
In this article I tell the story of how I came to codify my DL research setup and dive into the technical decisions that shaped it. I hope many of you will find it useful, whether you're looking for alternatives to Google Colab, having
2.
by aburla 4y ago
Any tools you would recommend for the scenario where you'd use monorepos?
3.
Ask HN: ML and Monorepos
2 points by aburla 4y ago|2 comments
4.
How to Solve the Model Serving Component of the MLOps Stack(alexandruburlacu.github.io)
1 points by aburla 4y ago|1 comments
5.
by aburla 4y ago
In this article I tried to summarize my personal thought process when deploying machine learning models. It's not quite a tutorial but can give you some good directions when faced with the problem of deciding how to serve a machine lea
6.
Going beyond simple error analysis of ML systems(alexandruburlacu.github.io)
1 points by aburla 4y ago|1 comments
7.
by aburla 4y ago
A few months ago I wrote an article about some considerations of how to perform error analysis for a trained machine learning model. It also proposes a “framework”/progression ladder for what methods to implement first. I’d be happy if
8.
by aburla 4y ago
I find having a simple baseline, and ideally a human baseline, extra important for a real world project. I had once the experience of working on a complicated project without any human baseline, even when requested. It was like walking thro
9.
by aburla 4y ago
Basically I assume a good AutoML solution, given time and enough compute, can come up with a very good model, much better than I can do in a few weeks, assuming no feature engineering. This way I can focus more on what could be missing and
10.
AutoML Solutions: What I Like and Don’t Like About AutoML as a Data Scientist(alexandruburlacu.github.io)
73 points by aburla 4y ago|10 comments