Amir Yazdan Profile
Amir Yazdan

@ayazdanb

Followers
2K
Following
12K
Media
47
Statuses
2K

Research Scientist at Google DeepMind. TPU Co-design, Machine Learning and Accelerator Design. #ComputerArchitecture #ML4Code

Mountain View, CA
Joined December 2017
Don't wanna be here? Send us removal request.
@ayazdanb
Amir Yazdan
4 days
If you’re only cheering breakthroughs from your own team, you’re missing the point. Celebrate good science everywhere, even from your rivals.
0
0
6
@ayazdanb
Amir Yazdan
5 days
RT @demishassabis: Official results are in - Gemini achieved gold-medal level in the International Mathematical Olympiad! 🏆 An advanced ver….
Tweet card summary image
deepmind.google
Our advanced model officially achieved a gold-medal level performance on problems from the International Mathematical Olympiad (IMO), the world’s most prestigious competition for young...
0
766
0
@ayazdanb
Amir Yazdan
6 days
RT @Yuchenj_UW: OpenAI and DeepMind models winning IMO golds is super cool, but not surprising if you remember AlphaGo beat Lee Sedol. Wha….
0
76
0
@ayazdanb
Amir Yazdan
9 days
Tweet media one
0
3K
0
@ayazdanb
Amir Yazdan
12 days
A more detailed look at our #ICML2025 work (led by @tjingrant and @ellieyhc) in collaboration with @GoogleDeepMind: PASTA .teaches LLMs, for the first time, to adaptively parallelize their own decoding strategy. No brittle heuristics. Just learned semantic independence.
@MIT_CSAIL
MIT CSAIL
12 days
A new approach from CSAIL & Google marks a shift toward teaching models to orchestrate their own parallel decoding strategy. The team's "Parallel Structure Annotation" (PASTA) enables LLMs to generate text in parallel, accelerating their response times:
Tweet media one
0
2
9
@ayazdanb
Amir Yazdan
15 days
You can dislike Elon, X, or dismiss Grok as irrelevant, but at the end of the day, science moves forward. Progress in AI should be welcomed, no matter where it comes from. Congrats to the xAI team!.
0
0
7
@ayazdanb
Amir Yazdan
16 days
A little extreme but seems on the right track….
@lazowska
Ed Lazowska
16 days
‘Coding is dead’: UW computer science program rethinks curriculum for the AI era via @GeekWire.
0
0
2
@ayazdanb
Amir Yazdan
21 days
5/5; So Happy Fourth. But let’s celebrate by holding this country to its highest ideals. Not just liberty and justice for some, but for all. That’s the America I want to grow old in!.
0
0
1
@ayazdanb
Amir Yazdan
21 days
4/5; Yet despite all this, I’m deeply grateful for the opportunities I’ve had here. I know I’ve benefited from privilege, and I want to see a future where those opportunities aren’t the exception, but the norm for everyone.
1
0
1
@ayazdanb
Amir Yazdan
21 days
3/5; We dropped nuclear bombs on civilians. We meddle in other nations’ democracies. And too often, we confuse nationalism with progress.
1
0
1
@ayazdanb
Amir Yazdan
21 days
2/5; We lead the world in innovation and opportunity, but also in wealth inequality, healthcare costs, student debt, and incarceration. Our immigration system is broken. Our foreign policy has harmed as much as it has helped.
1
0
1
@ayazdanb
Amir Yazdan
21 days
1/5; Happy Fourth of July 🇺🇸.America can be a great country, but is it great for everyone, or just the privileged? Today’s a good day to reflect: is this the best place to live for middle-class families, for immigrants, for the working poor?.
1
0
1
@ayazdanb
Amir Yazdan
22 days
There’s a troubling disconnect when researchers who are economically insulated downplay the risks AGI presents to vulnerable labor sectors.
0
0
3
@ayazdanb
Amir Yazdan
23 days
RT @RylanSchaeffer: New position paper! Machine Learning Conferences Should Establish a “Refutations and Critiques” Track. Joint w/ @sanmi….
0
54
0
@ayazdanb
Amir Yazdan
25 days
Excited to share the result of a fantastic collaboration between @berkeley_ai, @MIT_CSAIL, and @GoogleDeepMind! When BAIR collab goes brrr! .Check out this great summary, and find the lead authors at #ICML2025 to learn how to make LoRA fine-tuning faster and more efficient!.
@xiuyu_l
Xiuyu Li @ ICML
26 days
Sparsity can make your LoRA fine-tuning go brrr 💨. Announcing SparseLoRA (ICML 2025): up to 1.6-1.9x faster LLM fine-tuning (2.2x less FLOPs) via contextual sparsity, while maintaining performance on tasks like math, coding, chat, and ARC-AGI 🤯. 🧵1/.
0
0
15
@ayazdanb
Amir Yazdan
1 month
RT @henryHM_ko: I wrote a new blog on TPUs -- it's been fun seeing how different they are from GPUs and also drawing things on excalidraw a….
0
185
0
@ayazdanb
Amir Yazdan
1 month
RT @profvjreddi: Happy Father’s Day! 👨‍👧 This summer, my daughter Nora set a goal: to walk 1 million steps to support higher education. It’….
Tweet card summary image
mypacer.com
Join this walking challenge from Nora with me!
0
1
0
@ayazdanb
Amir Yazdan
1 month
RT @LeadersJunction: This 🔥
Tweet media one
0
2K
0
@ayazdanb
Amir Yazdan
1 month
RT @CompSciFact: 'Thinking doesn't guarantee that we won't make mistakes. But not thinking guarantees that we will.' -- Leslie Lamport.
0
21
0
@ayazdanb
Amir Yazdan
1 month
RT @bgurley: I actually don’t think there is any sign that “model quality” will be the key factor. If anything, the models all seem super c….
0
111
0