Director, Core AI, IBM. Chief Architect instructLAB.ai . Founder, Red Hat AI Innovation Team. PI @MITIBMLab. ❤️ Density Ratios.
Joined July 2009
- Tweets 643
- Following 1,179
- Followers 1,304
- Likes 2,864
44 Photos and videos
Akash Srivastava retweeted
5 Dec 2025
come check out poster #5518 at NeurIPS morning session today to learn about how you can encourage diversity / prevent early-pruning during inference-time scaling and boost the performance of any model without additional training!
6 Feb 2025
[1/x] can we scale small, open LMs to o1 level? Using classical probabilistic inference methods, YES! Joint @MIT_CSAIL / @RedHat AI Innovation Team work introduces a particle filtering approach to scaling inference w/o any training! check out probabilistic-inference-scal…
1
1
14
1,563
17 Jun 2025
What does it take to scale AI beyond the lab?
At #RedHatSummit, @ishapuri101 and I spoke with Red Hat CEO Matt Hicks & CTO Chris Wright on inference-time scaling, open infra (LLMD), and making AI affordable for enterprise.
🎧 youtu.be/mj1dwrPfvb4
#NoMathAI @RedHat_AI
1
8
5,503
24 Apr 2025
🚀 How is generative AI transforming the way we design cars, planes, and entire systems?
In Ep 2 of No Math AI, @ishapuri101 and I chat with Dr. @_faezahmed (@MIT DeCoDE Lab) about how AI boosts creativity, cuts design time, and works with engineers—not against them.
24 Apr 2025
How is generative AI reshaping engineering design?
In Episode 2 of No Math AI, hosts Dr. Akash Srivastava (@variational_i) and MIT PhD student Isha Puri (@ishapuri101) sit down with Dr. Faez Ahmed (@_faezahmed) from MIT DeCoDE Lab to explore just that. 👇
1
3
1,049
5 Apr 2025
SQuat: KV-Cache for making reasoning models go 🚀
📄paper: lnkd.in/emKhAVZu
💻 code: lnkd.in/e8TJ7N3R
From my awesome collaborators @RedHat_AI
5 Apr 2025
[1/x] 🚀 We're excited to share our latest work on improving inference-time efficiency for LLMs through KV cache quantization---a key step toward making long-context reasoning more scalable and memory-efficient.
2
10
1,246
Akash Srivastava retweeted
2 Apr 2025
Excited to share our preliminary work on customizing reasoning models using Red Hat AI Innovation’s Synthetic Data Generation (SDG) package!
📄 Turn your documents into training data for LLMs.
🧵👇
2
5
10
1,322
Akash Srivastava retweeted
4 Mar 2025
had a great time giving a talk about probabilistic inference scaling and the power of small models at the IBM Research ML Seminar Series - the best talks end with tons of questions, and it was great to see everyone so engaged : ) youtube.com/watch?v=--3rsQwM…
2
21
140
14,965
7 Feb 2025
Come along and help us build reasoning in small LLMs
🚀 Exploring LLM reasoning—live!
We, the @RedHat AI Innovation Team, are working on reproducing R1-like reasoning in small LLMs without distilling R1 or its derivatives.
We’re documenting our journey in real-time:
🔗 Follow along: red-hat-ai-innovation-team.g…
3
435
6 Feb 2025
Excited to share our latest work with @ishapuri101 et al.! 🚀 We introduce a probabilistic inference approach for inference-time scaling of LLMs using particle-based Monte Carlo methods—achieving 4–16x better scaling on math reasoning tasks and O1-level performance on MATH500.
6 Feb 2025
[1/x] can we scale small, open LMs to o1 level? Using classical probabilistic inference methods, YES! Joint @MIT_CSAIL / @RedHat AI Innovation Team work introduces a particle filtering approach to scaling inference w/o any training! check out probabilistic-inference-scal…
5
427
Akash Srivastava retweeted
18 Dec 2024
🧩 Why do task vectors exist in pretrained LLMs?
Our new research uncovers how transformers form internal abstractions and the mechanisms behind in-context learning(ICL).
6
30
188
21,909
Akash Srivastava retweeted
17 Sep 2024
Neural activity is correlated among animals performing the same task and across sequential trials.
Led by @zhang_yizi and @hl3616, we develop an reduced-rank model that exploits shared structure across animals to improve neural decoding.
biorxiv.org/content/10.1101/…
1
35
189
14,808
Akash Srivastava retweeted
24 Jul 2024
What will a foundation model for the brain look like?
We argue that it must be able to solve a diverse set of tasks across multiple brain regions and animals.
Check out our preprint where we introduce a multi-region, multi-animal, multi-task model (MtM): arxiv.org/abs/2407.14668
5
62
254
36,469
Akash Srivastava retweeted
14 May 2024
🚀 Stronger, simpler, and better! 🚀
Introducing Value Augmented Sampling (VAS) - our new algorithm for LLM alignment and personalization that outperforms existing methods!
4
33
128
25,534
Akash Srivastava retweeted
11 May 2024
Excited to give a talk on our hottest, newest work “Value Augmented Sampling for Language Model Alignment and Personalization” at 2:30p Halle A3 in #ICLR2024 Reliable and Responsible Foundation Models Workshop 🥳🥳
11 May 2024
📢Workshop on Reliable and Responsible Foundation Models will happen today (8:50am - 5:00pm). Join us at #ICLR2024 room Halle A 3 for a wonderful lineup of speakers, along with 63 amazing posters and 4 contributed talks! Schedule: iclr-r2fm.github.io/#program.
1
2
12
1,470
5 May 2024
Attending #ICLR2024, interested in continual learning and like probabilistic modeling? Lazar from the @MITIBMLab, will be presenting our latest work that takes a probabilistic approach to modular continual learning on Tuesday, 7 May, Halle B #222 (iclr.cc/virtual/2024/poster/…).
5 May 2024
I’ll be presenting our #ICLR2024 paper on a probabilistic approach to scaling modular continual learning algorithms while achieving different types of knowledge transfer. (arxiv.org/abs/2306.06545, in collaboration with @variational_i @swarat @RandomlyWalking ). A tldr (1/8):
1
11
1,094
Akash Srivastava retweeted
13 Apr 2024
Check out our work titled "From Automation to Augmentation: Redefining Engineering Design and Manufacturing in the Age of NextGen-AI", where we highlight the requirements for NextGenAI suitable for design, engineering, and manufacturing.
mit-genai.pubpub.org/pub/9s6…
Instead of continuing to emphasize automation, a human-centric approach to the next generation of #AI technologies in #manufacturing could enhance workers' skills and boost productivity.
mit-genai.pubpub.org/pub/9s6…
@AustinLentsch @DAcemogluMIT @baselinescene @_faezahmed @MITMechE
1
3
13
3,084
6 Mar 2024
New work from @MITIBMLab researchers on large scale alignment of LLMs. Check out the models at HF huggingface.co/ibm/merlinite…
6 Mar 2024
Hey, we did a thing: "LAB: Large-scale Alignment for chatBots"—a new synthetic data-driven LLM alignment method that yields great results without using large-scale human or proprietary model data.
arxiv.org/abs/2403.01081
models: huggingface.co/ibm/labradori…, huggingface.co/ibm/merlinite…
1
6
462
5 Mar 2024
New work on automated red-teaming in LLMs using curiosity-driven exploration! #iclr24
5 Mar 2024
(1/4) 🎉 Excited to share our ICLR'24 paper on "Curiosity-driven Red-teaming for Large Language Models"! We bridge curiosity-driven exploration in reinforcement learning (RL) with red-teaming, introducing the Curiosity-driven Red-teaming (CRT) method. #ICLR24 #AI #LLMSecurity
1
13
925
15 Dec 2023
❤️ #NeurIPS2023. After 4 years, met my adviser and my adviser's advisor at the same time.
3
3
112
16,076
Akash Srivastava retweeted
11 Dec 2023
Uniform sampling hampers offline RL. How to fix it? Check our paper at #NeurIPS2023.
Time: Wed 13 Dec 5 p.m. CST — 7 p.m. CST
Location: Great Hall & Hall B1 B2 (level 1) #1908
Paper: openreview.net/forum?id=TW99…
Code: github.com/Improbable-AI/dw-…
1
1
4
630