@vboykis@panozzaj With a model? If it’s a short timeline I might go with a simpler version that works good enough and iterate on it (test out other models) when v0 is out n/n
@vboykis@panozzaj Is my data low or high dimension? Do I think a lot of features matter? Is it a text or image problem? How long and what is the cost to train it? What is the latency when serving the prediction? (How does it scale?) also an important thing is how long do you have to come up 2/n
The women you see today in tech - we have made sacrifices to follow our passion and be in this field. Kudos to all of us that have stuck it out and are becoming who we want to be - we have generations looking up to us, we are blazing a path. 🔥
[paid] Summer internship opportunities at @nygenome:
https://t.co/UAA20aazXO
See in particular those in the Bioinformatics team:
https://t.co/YW7TzY8JbM
Please RT
New blog post: I have been pretty unhappy with the state of modern development and wanted to understand why. We are building on millions of dependencies we don’t understand, in fragmented and proprietary cloud environments we can’t introspect.
https://t.co/tosBNom9Bu
With all the layoff and crazy things happening, I decided to open my calendar again in Nov for 1:1 with me (30 minutes during lunch). You can ask me anything about DS/ ML career, life at Netflix, reviewing your resume, talk about a Netflix position you want to apply to etc.1/n
🤯 Users often ask: “Which test to choose for data drift detection?”
👩🏻🔬The @EvidentlyAI team ran an experiment to help build intuition on how different statistical tests behave.
We compared five popular ML drift detection methods.
Here is what we did 👇🏼
(1/4)
I am super excited to share that I joined @NetflixEng as a senior machine learning data scientist. I’ll be part of the content - creative excellence team and I can’t wait to work on that!
The vocabulary we have around burn out is insufficient.
It's often framed as too much work (which is definitely one factor), but I see it also stem from misalignment in terms of strategy or values, lack of recognition, and/or sense of belonging.
Very good thread! When I started working in ML (with a stat background), I was so confused because people would use the term inference to talk about prediction!
Yesterday I asked twitter about the difference between inference and prediction. I’d like to make sense of a few things in a 🧵
1/
https://t.co/XXEMlgIICG
them friends now. I am sad to leave such an amazing environment but I am very excited about my next step!
But first I am taking a break until mid April to settle in our new house and get some quality time with my daughter and my 4 months old :) 2/2
Today is bittersweet as it was my last day at Zillow. I am very proud of everything I accomplished during those nearly 3 years. But more than the work itself what I will miss the most is the people. I made some amazing connections there and I am lucky enough to call some of 1/2