Prateek Jain (@jainprateek_)

Prateek Jain Retweeted

H

Research Engineer, Tokyo: job-boards.greenhouse.io/deepmind/jobs/… Research Scientist, Tokyo: job-boards.greenhouse.io/deepmind/jobs/… Research Scientist, Bangalore: job-boards.greenhouse.io/deepmind/jobs/…

1

11

30

9

5.0K

P

Prateek Jain@jainprateek_ · Jul 23

We're hiring @GoogleDeepMind APAC! 🇯🇵🇮🇳 Seeking experts in multilingual, multicultural & multimodal AI to help shape the future of our frontier models including Gemini. This is a unique opportunity to make impacts on billions of users.

PPartha Talukdar (✈️ ACL 25)@partha_p_t · Jul 23

@GoogleDeepMind India 🇮🇳 & Japan 🇯🇵 are looking for strong candidates in multilinguality, multicultural, & multimodality areas. RS Bangalore: job-boards.greenhouse.io/deepmind/jobs/… RS Tokyo: job-boards.greenhouse.io/deepmind/jobs/… RE Tokyo: job-boards.greenhouse.io/deepmind/jobs/…

12

59

527

132

58.0K

Prateek Jain Retweeted

P

Partha Talukdar (✈️ ACL 25)@partha_p_t · Jul 23

@GoogleDeepMind India 🇮🇳 & Japan 🇯🇵 are looking for strong candidates in multilinguality, multicultural, & multimodality areas. RS Bangalore: job-boards.greenhouse.io/deepmind/jobs/… RS Tokyo: job-boards.greenhouse.io/deepmind/jobs/… RE Tokyo: job-boards.greenhouse.io/deepmind/jobs/…

2

24

148

77

64.0K

P

Prateek Jain@jainprateek_ · Jul 15

🪆 Matryoshka is extremely general & applicable to every component in our modern ML/DL stack. It can't get more fundamental that 🪆 in bit space to enable elastic quantization! Drop by the poster and say hi to Puranjay (on behalf of @pranavn1008 @JeffDean @jainprateek_ & me).

PPURANJAY DATTA@puranjay1412 · Jul 15

Hi, I'll be presenting Matryoshka Quantization (arxiv.org/abs/2502.06786) on 16th July at #ICML2025 📍East Exhibition Hall A-B #3606 ⏲️ 11 AM - 1:30 PM

1

9

61

16

5.0K

P

Prateek Jain@jainprateek_ · Jul 15

Check out our #ICML2025 poster on Matryoshka Quantization tomorrow, being presented by @puranjay1412 on behalf of all the authors.

PPURANJAY DATTA@puranjay1412 · Jul 15

Hi, I'll be presenting Matryoshka Quantization (arxiv.org/abs/2502.06786) on 16th July at #ICML2025 📍East Exhibition Hall A-B #3606 ⏲️ 11 AM - 1:30 PM

7

22

167

28

30.0K

P

Prateek Jain@jainprateek_ · Jul 15

Nested matformer style models for faster parallel decoding! Sahil will present this poster at ICML today. Please check this one out if you are interested in architecture, efficiency etc. Additionally, Sahil will be applying for grad schools this cycle. He is brilliant, has a…

SSahil Goyal@sahilgo6801 · Jul 12

Hi, we'll be presenting MaGNeTS (arxiv.org/pdf/2502.00382) on 15th July at #ICML2025 📍East Exhibition Hall A-B #3209 🕦 11 AM - 1:30PM Excited to discuss about nested transformers and decode time scaling for visual generation!

0

2

33

5

3.0K

P

Prateek Jain@jainprateek_ · Jul 15

Puranjay will present our poster on nested bitwise models or MatQuant, so if you are ICML and interested in the topic, do bother him :) Puranjay is going on the grad-school market this cycle. So if you are looking for a brilliant, hardworking student with good ML+LLM exposure,…

PPURANJAY DATTA@puranjay1412 · Jul 15

Hi, I'll be presenting Matryoshka Quantization (arxiv.org/abs/2502.06786) on 16th July at #ICML2025 📍East Exhibition Hall A-B #3606 ⏲️ 11 AM - 1:30 PM

0

4

42

6

20.0K

P

Prateek Jain@jainprateek_ · Jul 14

Powerful gemini embeddings with support for flexible output dimensions 🪆🪆🪆

GGoogle AI Developers@googleaidevs · Jul 14

📢 The Gemini Embedding text model (gemini-embedding-001) is now generally available in the Gemini API via Google AI Studio. It supports 100+ languages and uses Matryoshka Representation Learning for flexible output dimensions, allowing devs to scale down from 3072 dimensions.

0

1

19

2

1.0K

P

Prateek Jain@jainprateek_ · Jul 14

Embeddings are an important part of the magic of deep learning, allowing diverse data to be represented semantically as compact vectors. These embedding vectors can be used to retrieve, compare and classify data for downstream tasks. Gemini Embeddings model is now available!

PPhilipp Schmid@_philschmid · Jul 14

Gemini Embeddings General Available! First Gemini Embedding model (`gemini-embedding-001`) now available for production use, ranking top on MMTEB optimized for finance, science, legal, search, code. 👀 - 🏦 $0.15 per million input tokens with free tier. - 🥇 Top MMTEB…

1

4

39

16

4.0K

Prateek Jain Retweeted

P

Philipp Schmid@_philschmid · Jul 14

Gemini Embeddings General Available! First Gemini Embedding model (`gemini-embedding-001`) now available for production use, ranking top on MMTEB optimized for finance, science, legal, search, code. 👀 - 🏦 $0.15 per million input tokens with free tier. - 🥇 Top MMTEB…

30

101

819

451

77.0K

P

Prateek Jain@jainprateek_ · Jul 12

Hi, we'll be presenting MaGNeTS (arxiv.org/pdf/2502.00382) on 15th July at #ICML2025 📍East Exhibition Hall A-B #3209 🕦 11 AM - 1:30PM Excited to discuss about nested transformers and decode time scaling for visual generation!

SSahil Goyal@sahilgo6801 · May 1

My first work @GoogleDeepMind accepted to ICML!

1

12

84

28

12.0K

Prateek Jain Retweeted

J

Jaydev Tonde@JaydevTonde · Jul 2

Google DeepMind just launched #Gemma3n! 🚀 They've released E2B & E4B, with more sizes coming for your specific hardware needs. What's mind-blowing? You can extract new, custom models from the bigger one.

1

2

11

4

1.0K

P

Prateek Jain@jainprateek_ · Jul 1

My team is hiring a Technical Program Manager to help organize, accelerate, and empower world-class research. High-impact, high-growth role for someone passionate about AI and great at making things happen.

PPrateek Jain@jainprateek_ · Jun 30

We are hiring Technical Program Manager to organize and enable our research teams to be the best at what they do and to make fast-paced progress towards our mission of bringing AGI responsibly. Ideal candidates should have a demonstrable record of strong program management…

0

3

28

6

5.0K

P

Prateek Jain@jainprateek_ · Jun 30

We are hiring Technical Program Manager to organize and enable our research teams to be the best at what they do and to make fast-paced progress towards our mission of bringing AGI responsibly. Ideal candidates should have a demonstrable record of strong program management…

1

6

44

18

15.0K

P

Prateek Jain@jainprateek_ · Jun 28

Similiar to how virtualization enabled elasticity of compute in Cloud, which led to huge adoption of Cloud, I have felt that Matryoshka architectures, devised by @jainprateek_, @adityakusupati & team are enabling elasticity of FM capability. People will recognize it over time.

AAditya Timmaraju@tadityasrinivas · Jun 27

Great to see the power of Matryoshka🪆 architectures, devised by my lab, getting widely highlighted as an indispensable component of the LLM cognitive core. cc @jainprateek_ @adityakusupati A sample list of Matryoshka papers from my lab: Matryoshka Representation Learning…

0

3

26

4

3.0K

P

Prateek Jain@jainprateek_ · Jun 27

The race for LLM "cognitive core" - a few billion param model that maximally sacrifices encyclopedic knowledge for capability. It lives always-on and by default on every computer as the kernel of LLM personal computing. Its features are slowly crystalizing: - Natively multimodal…

OOmar Sanseviero@osanseviero · Jun 26

I’m so excited to announce Gemma 3n is here! 🎉 🔊Multimodal (text/audio/image/video) understanding 🤯Runs with as little as 2GB of RAM 🏆First model under 10B with @lmarena_ai score of 1300+ Available now on @huggingface, @kaggle, llama.cpp, ai.dev, and more

383

1.0K

10.0K

5.0K

1.2M

P

Prateek Jain@jainprateek_ · Jun 27

Great to see the power of Matryoshka🪆 architectures, devised by my lab, getting widely highlighted as an indispensable component of the LLM cognitive core. cc @jainprateek_ @adityakusupati A sample list of Matryoshka papers from my lab: Matryoshka Representation Learning…

AAndrej Karpathy@karpathy · Jun 27

The race for LLM "cognitive core" - a few billion param model that maximally sacrifices encyclopedic knowledge for capability. It lives always-on and by default on every computer as the kernel of LLM personal computing. Its features are slowly crystalizing: - Natively multimodal…

0

2

38

28

8.0K

P

Prateek Jain@jainprateek_ · Jun 27

It is great to see my first contributed project at Google, Gemma3n, is now fully open-sourced after Google I/O Preview annoucement! 🎉 Learned so much while contributing to parts of Gemma3n alongside amazing colleagues. We did a lot for both model quality and efficiency!

GGoogle DeepMind@GoogleDeepMind · Jun 26

We’re fully releasing Gemma 3n, which brings powerful multimodal AI capabilities to edge devices. 🛠️ Here’s a snapshot of its innovations 🧵

1

6

41

5

4.0K

Prateek Jain Retweeted

X

Xuan-Son Nguyen@ngxson · Jun 26

Gemma 3n has arrived in llama.cpp 👨‍🍳 🍰 Comes in 2 flavors: E2B and E4B (E means "effective/active parameters")

3

5

80

20

4.0K