Shreya Shankar
PhD student in databases focusing on MLOps @Berkeley_EECS @UCBEPIC. Prev. 1st ML engineer @viaduct_ai, research @googlebrain, BS & MS @Stanford CS. She/they.
View on πThreads
Of course prompt engineering is a job. ML systems and software have always been an amalgamation of models and rules. Now, with GPT et al., it is really hard to figure out what rule...
I want to talk about my data validation for ML journey, and where Iβm at now. I have been thinking about this for 6 ish years. It starts with me as an intern at FB. The task was to...
Iβve been talking a lot at the @UCBEPIC events about the need for a framework that allows people to build hybrid rule-based and AI-based systems. Current solutions force people to...
Only in this past year did I realize how much free labor I was giving away. Iβm still upset about it, and Iβm also upset that people disproportionately take advantage of my women &...
Ok I have thoughts on whether there should be separate CS courses for those who want to just use the tool vs go in deep. 5 years ago I thought absolutely, why learn about semaphore...
I realize I didn't do a good job discussing this. Here's an example of the KS false positive problem in the ML pipeline setting: suppose I have 1 feature, age, computed from differ...
I think it's great that big model cos are both (1) trying to think about MLOps and (2) providing me with more outlets to speculate about the future βΊοΈ https://t.co/Qjbb0KRicn
Our understanding of MLOps is limited to a fragmented landscape of thought pieces, startup landing pages, & press releases. So we did interview study of ML engineers to understand...
Yes, this. It is not really about public vs private schools. I resisted commenting for a day but now I have to. I don't even know how to compare Stanford & Berkeley CS https://t.co...
Unit testing for ML is a big category of questions but here are my thoughts on the data validation piece (ensuring model inputs/outputs have good "quality" such that ML performance...
Sunday morning thoughtdump of some reasons why I believe "doing ML in the DBMS" doesn't make sense for traditional OLTP DB architectures (sorry, I'm studying for my prelim and this...
I'm excited (and nervous) to post this thread: I've always known I wanted a partner but didn't know what a supportive one looked like (esp. as an ambitious woman who wants kids som...