Praneetha Vaddamanu

Applied Scientist at Microsoft Turing, working on LLM reasoning agents. @CarnegieMellon @bitspilaniindia

[Praneetha Vaddamanu @pranurs]

Presenting our work at the SCALR workshop @ COLM 2025, Montreal tomorrow! Stop by for a chat if you're interested :D

Paper: https://t.co/UvGHMaBr71

[Sanchit Ahuja @SanchitAhuja7]

This got accepted at SCALR workshop at COLM 2025 AND EMNLP Findings 2025!
@pranurs will be presenting this work at COLM and hopefully I at EMNLP. :)

[Praneetha Vaddamanu @pranurs]

@_jasonwei 🥺

[Praneetha Vaddamanu @pranurs]

Excited to share our new paper on cross-lingual LLM reasoning - with @SanchitAhuja7 and Barun!

Turns out models may reason more efficiently in langs like Arabic or Korean- cutting tokens without hurting accuracy. A step toward rethinking the default role of English in reasoning!

[Sanchit Ahuja @SanchitAhuja7]

New paper!
Can reasoning in non-English languages be token-efficient and accurate?
We evaluate this across 3 models, 7 languages, and 4 math benchmarks.
Here’s what we found 🧵 (1/n)

[Kevin Meng @mengk20]

why do language models think 9.11 > 9.9?

at @transluceAI we stumbled upon a surprisingly simple explanation - and a bugfix that doesn't use any re-training or prompting.

turns out, it's about months, dates, September 11th, and... the Bible?

[Yifei He @heyifei99]

Wonder how to efficiently balance languages in your multilingual LLM training? Curious about how those ratios generalize across model sizes?

📢Thrilled to share my @Microsoft internship work on scaling laws for multilingual LMs! For multilingual pretraining, we unlock a power-law relationship between loss, model size, dataset size, and most importantly, language sampling ratios. This allows us to derive **optimal sampling ratios** that strike the best balance of languages at different scales!

Preprint: https://t.co/peaWiDMjZ7

🧵👇

[Praneetha Vaddamanu @pranurs]

@RafaelNadal Will miss you on court Rafa ❤️ thanks for being such an inspiration 👏

[Wonder of Science @wonderofscience]

"The size and age of the Cosmos are beyond ordinary human understanding. Lost somewhere between immensity and eternity is our tiny planetary home."
— Carl Sagan

Zoom out from Earth to the edge the Universe.

Animation Credit: ESO/L.Calçada. Music: Monolake

[Praneetha Vaddamanu @pranurs]

@WorldAndScience @nihaljain98

[Android Developers @AndroidDev]

🥁🥁🥁🥁🥁🥁🥁🥁🥁🥁🥁🥁🥁

We challenged you to build helpful on-device machine learning apps and the results are in!

The 10 winners for the #AndroidDevChallenge are …

Read more → https://t.co/iEsWuxOxNd
Download the apps → https://t.co/ZQgHFUdZE2

#11WeeksOfAndroid