Robert Nishihara Profile
Robert Nishihara

@robertnishihara

Followers
13K
Following
4K
Media
138
Statuses
2K

Co-founder @anyscalecompute. Co-creator of @raydistributed. Previously PhD ML at Berkeley.

Joined March 2009
Don't wanna be here? Send us removal request.
@robertnishihara
Robert Nishihara
5 months
Beyond pre-training, here's how I imagine most learning will work. 1. AI models / systems will maintain large collections of retrievable knowledge. This will include facts like "the capital of California is Sacramento" and tactics like "when playing Monopoly, buy a bunch of
@robertnishihara
Robert Nishihara
5 months
We're missing techniques for "training-time reasoning." Right now there's a lot of progress on inference-time reasoning, which is incredibly cool (I use o3 all the time). If I think about how I learn stuff, e.g., when reading a technical paper, it's very compute intensive. Most
4
15
141
This was one of the highlights of Ray Summit last week.
@srush_nlp
Sasha Rush
2 days
Talk at Ray Summit on "Building Cursor Composer." Overview of the work from our research team. https://t.co/9a5yeC3IT8
1
2
15
Ray Summit 2025 just wrapped, and what an incredible week it's been! Thank you all for coming! Thousands of builders, researchers, and platform engineers came together in San Francisco to share how they’re scaling AI in production with Ray.
1
2
10
@srush_nlp
Sasha Rush
2 days
The Ray community is quite impressive. RL requires rethinking much of the distributed story of ML. Enjoyed hearing about all the different projects looking to improve these approaches, particularly open source projects like verl and SkyRL.
2
6
46
@srush_nlp
Sasha Rush
2 days
Talk at Ray Summit on "Building Cursor Composer." Overview of the work from our research team. https://t.co/9a5yeC3IT8
7
43
327
@PyTorch
PyTorch
13 days
Anyscale has contributed Ray to #PyTorch Foundation, joining #PyTorch, vLLM, and DeepSpeed under open governance to advance scalable, distributed AI infrastructure. @janakiramm (@Forbes) on how this model reduces fragmentation and fosters transparent collaboration ⬇️
4
11
64
@matthew_d_white
Matt White
5 days
Just wrapped an incredible week at #RaySummit2025! Great insights from @DrJimFan (NVIDIA) on embodied AI & Project GR00T, and an inspiring panel with @JosephSpisak, @SimonMo_AI, Dawn Chen & @robertnishihara. Proud to see @anyscaleinc contribute #Ray to @PyTorch Foundation
0
1
5
@d_spodarets
Dmytro Spodarets
7 days
I had a great conversation with Keerti Melkote, CEO of Anyscale, at Ray Summit 2025 here in San Francisco. We talked about Ray’s new chapter within the Linux Foundation, Anyscale’s partnership with Microsoft, and the future of AI infrastructure. It was a truly engaging and
0
1
2
@SiliconANGLE
SiliconANGLE
7 days
Anyscale and Microsoft team up to launch AI-native compute service on Azure
Tweet card summary image
siliconangle.com
Anyscale and Microsoft team up to launch AI-native compute service on Azure - SiliconANGLE
0
1
2
@intelcapital
Intel Capital
6 days
A new #partnership between @anyscalecompute and @Microsoft will deliver a new AI-native #compute service as a first-party offering on Azure, offering a high-performance experience for scaling up advanced #AI workloads. More in @SiliconANGLE from @duncanriley:
Tweet card summary image
siliconangle.com
Anyscale and Microsoft team up to launch AI-native compute service on Azure - SiliconANGLE
@anyscalecompute
Anyscale
7 days
We’re excited to announce Anyscale on Azure, a new AI-native compute service co-engineered with Microsoft delivered as a fully managed, first-party offering on Azure. Now in private preview.
0
3
12
@satyampsoni
Satyamm Soni
6 days
I am here at the Ray AI Summit in SF, the conversations and energy is incredible.
0
2
20
@aravind646
Aravind Selvan
6 days
Caught the RL @cursor_ai session at @anyscalecompute’s Ray Summit yesterday. Wild to see how much infra goes into making Composer, their interactive coding model, actually work: Firecracker sandboxes, Ray actors everywhere, semantic search harnesses tied to training. The future
0
3
7
@talkdevnow
TalkDev
7 days
@anyscalecompute and Microsoft have partnered to launch an AI-native compute service on Azure Kubernetes Service (AKS), powered by the open-source framework Ray. Anyscale+2Anyscale+2 👉 Read the full story: https://t.co/nmGyEPCKkl #AI #Cloud #Azure #EnterpriseTech #Innovation
Tweet card summary image
talkdev.com
Anyscale and Microsoft launch a private preview of a new AI-native compute service on Azure, designed for efficient management and scaling of AI workloads in enterprises.
0
2
2
@robertnishihara
Robert Nishihara
6 days
Yesterday at Ray Summit, we announced a partnership between @anyscalecompute and @Azure! This brings Anyscale's managed & optimized @raydistributed experience to the Azure ecosystem. Anyscale on Azure is now in private preview and is accessible directly through the Azure Portal.
Tweet card summary image
devblogs.microsoft.com
The path from prototype to production for AI/ML workloads is rarely straightforward. As data pipelines expand and model complexity grows, teams can find
1
7
39
@jfiance
Jeremy Fiance
6 days
Ray Summit 2025 has been exceptional. This is where 1,000+ building production AI infrastructure come together. Not just talking about the future. Building it. Open source contributors, ML engineers running systems at scale, and infrastructure leaders sharing real solutions to
1
2
15
@NVIDIAAIDev
NVIDIA AI Developer
7 days
We’re excited to sponsor @anyscalecompute's #RaySummit ⭐ Today, Dr. Jim Fan took part in the keynote speech, and we connected with so many talented AI/ML builders to share hands-on insights. Explore how the NVIDIA Developer Program is driving open-source innovation and
10
14
63
@josh_wills
JosH100
7 days
Great talk on @datologyai trillion token synthetic data pipeline by @hurrycane and Fan Pan at #raysummit!
8
8
67
@LambdaAPI
Lambda
7 days
Scaling AI from prototyping to production can be hard: distributed jobs, reliability, cost. Ray makes it practical. Lambda makes it run. Read more → https://t.co/iHrVqGmg62 Meet us at @anyscalecompute's Ray Summit, booth S2.
0
1
4
@tyler_griggs_
Tyler Griggs
7 days
SkyRL tx talk starting soon (4pm) at the Ray Summit, come join if you’re around!
@tyler_griggs_
Tyler Griggs
8 days
SkyRL tx is now bumped to v0.1, which adds support for running @thinkymachines Tinker Cookbook RL loops unmodified out of the box! We'll be talking more about tx at Ray Summit tomorrow at 4pm at the Ray Summit, please join if you're around https://t.co/WNgonI2Z0W
0
2
20