Rohit Pgarg @rohitpgarg

I tweet about Deep Learning/ML, and tech in general. Working as Data Scientist on recommendation systems at @glancescreen and @Roposolove. Tweets are my own. Joined April 2016

Tweets

1K
Followers

75
Following

863
Likes

548

Jay @thetrickytrade

6 months ago

One of the most underrated yet effective strategies for snagging premium cabin awards? Knowing EXACTLY when airlines release their seats. Let me show you how this works, using Singapore Airlines as an example... 🧵👇 🎁 DM for bonus content at the end.

66 81 150 31K 241

Download Image

Rohit Pgarg @rohitpgarg

6 months ago

.@Swiggy @SwiggyCares You have stolen Rs 1532 for order 201939007131119. Did not give an option to cancel within 60 secs. Then customer support charged me the full amount for cancellation. Either revert the absurd cancellation charge, or give me my product back.

2 0 0 34 0

Rohit Pgarg @rohitpgarg

7 months ago

.@ImagineApplePR Where's my order # 75050148411? It's been a week and no response from your side. Pathetic.

2 0 0 12 0

Nick Dobos @NickADobos

10 months ago

Lmfao grok is worse at everything

AmebaGPT @amebagpt

10 months ago

Lmfao grok is worse at everything https://t.co/LieetqC5dH

11 18 200 105K 110

Download Image

35 18 442 82K 66

Download Image

Rohit Pgarg @rohitpgarg

2 years ago

It’s been almost 24 hours, and still no response. Pathetic work, @zomato @zomatocare

Rohit Pgarg @rohitpgarg

2 years ago

It’s been almost 24 hours, and still no response. Pathetic work, @zomato @zomatocare

1 0 0 83 0

1 0 0 64 0

Andrej Karpathy @karpathy

2 years ago

Random note on k-Nearest Neighbor lookups on embeddings: in my experience much better results can be obtained by training SVMs instead. Not too widely known. Short example: github.com/karpathy/rando… Works because SVM ranking considers the unique aspects of your query w.r.t. data.

106 489 4K 1.0M 3K

Rohit Pgarg @rohitpgarg

3 years ago

You haven’t learnt anything in your life.

Mark Tenenholtz @marktenenholtz

3 years ago

You haven’t learnt anything in your life.

180 33 398 186K 41

0 0 0 40 0

Radek Osmulski @radekosmulski

4 years ago

Curious about recommender models? Interested in endowing models from other domains with some of their superpowers? Please join me on a whirlwind tour of 6 recsys architectures! >> a thread 🧵 <<

8 122 685 0 410

Download Image

Rohit Pgarg @rohitpgarg

4 years ago

ImageNet is the new MNIST.

Aleksander Madry @aleks_madry

4 years ago

ImageNet is the new MNIST.

29 372 2K 0 462

Download Image

0 0 3 0 0

muesli 🐧 @mueslix

4 years ago

The Alternative Big O notation: O(1) = O(yeah) O(log n) = O(nice) O(n) = O(k) O(n²) = O(my) O(2ⁿ) = O(no) O(n!) = O(mg!) O(n!*n) = O(rly?)

35 2K 7K 0 385

Tanishq Mathew Abraham, Ph.D. @iScienceLuvr

4 years ago

Annotated PyTorch Paper Implementations by @labmlai is an AMAZING resource: • Deep learning papers explained in-depth with code side-by-side • Constantly updated with some of the latest papers! • 100% free and open-source! Check it out here → nn.labml.ai

4 225 1K 0 515

Download Image

Rohit Pgarg @rohitpgarg

4 years ago

Hey @fernsnpetals, suffering from poor support. Order delivered partially with no rectification. No refund, no re-delivery. No response.

2 0 3 0 0

Rohit Pgarg @rohitpgarg

4 years ago

Fourier features work in NLP too! Uber had a paper on DCT based features could replace lower layers of Resnet for image classification tasks. (limited) return of non-learned features!

Alexander Kolesnikov @kolesnikov

4 years ago

Fourier features work in NLP too! Uber had a paper on DCT based features could replace lower layers of Resnet for image classification tasks. (limited) return of non-learned features!

1 21 113 0 27

Download Image

0 0 1 0 0

Felice Antonio Merra @merrafelice

5 years ago

Very excited to share ELLIOT a Python Framework for #Reproducible #RecommenderSystems with more than 50 models, 36 metrics and many features to support easy-to-run experiments GitHub page: github.com/sisinflab/elli… #RecSys2021 #WSDM2021 #CIKM2021 #SIGIR202 #ECIR2021 (1/3)

2 12 60 0 14

Download Image

Eugene Yan @eugeneyan

5 years ago

How does real-time ML look in practice? We discuss real-time recommenders, drawing on my experience & industry papers: • When does it make sense? When does it not? • How have China & US companies implemented them? • How do we design & build an MVP? eugeneyan.com/writing/real-t…

9 41 160 0 56

Rohit Pgarg @rohitpgarg

5 years ago

reuters.com/article/uk-int… "In June, it lost one of its veteran chip designers, Jim Keller, over a dispute on whether the company should outsource more of its production, sources said at the time." Interesting.

0 0 0 0 0

Shreya Shankar @sh_reya

5 years ago

Unit testing for ML pipelines is challenging given changing data, features, models, etc. Changing I/O make it hard to have fixed unit tests. To hackily get around this, I liberally use assert statements in scheduled tasks. These have saved me so many times. Thread: (1/11)

15 128 825 0 235

morgan @CasualEffects

5 years ago

@AmazingThew Something life changing I learned early at NVIDIA is that "putting something in hardware" (creating an application-specific integrated circuit, ASIC) usually does not make a program faster. (thread):