Heading to MLSys in Bellevue? Come hang with LMSYS and RadixArk people!
RadixArk CTO @BanghuaZ will give the MLSys opening remarks on Day 1 and co-host the Young Professional Symposium talks and panels.
We're also co-hosting two happy hours (event links in the comments):
Mon, May 18
Happy hour with @allen_ai, sponsored by @CrusoeAI and @Doubleword_. Come meet @natolambert, @GenAI_is_real, @BanghuaZ, Connor Guerrero, Jamie Dborin, Carlo Mussolini, and the SGLang core members.
Tue, May 19
MLSys Happy Hour with @RadixArk, @EssenceVenture, and @DeltaInstitutes.
One more thing! If you'd like to sit down with someone from the team for research, partnerships, open source, or hiring, drop your details and we'll set up a 15 or 30 min chat: bit.ly/43kuPFz
See you in Bellevue!
🍻 MLSys 2026 Happy Hour is coming up!
SGLang & Ai2 (@allen_ai) are co-hosting a happy hour for the open AI community during MLSys 2026 in Bellevue, sponsored by @CrusoeAI & @Doubleword_!
Come hang with:
-@natolambert, Sr. Research Scientist at @allen_ai
-@GenAI_is_real, SGLang Core Dev
-@BanghuaZ, Co-founder of @radixark
-Connor Guerrero, Sr. DevRel Manager at @CrusoeAI
-Jamie Dborin, Co-founder & Head of Research at @Doubleword_ Inference Lab
-Carlo Mussolini, Member of Technical Staff at Fractile
Spots are limited, RSVP now and we'll see you there!
🕐 Mon, May 18 in Bellevue downtown
👉 RSVP required: luma.com/jdix71up
We’ve partnered with @Doubleword_ to bring our structured generation engine directly to their inference platform.
No more bad outputs!
👉 app.doubleword.ai
🩻🩻🩻 announcing SynthVision! 🚀🚀🚀
a new dataset from OpenMed x @huggingface x @Doubleword_
110k synthetic VQA dataset for medical records (2 order of magnitude size improvement on previous datasets)
don't let your medical model sleep on this data
@UnslothAI If anyone wants to run Nemotron on real workloads, we've made Nemotron 3 Super free during GTC on Doubleword - check it out at app.doubleword.ai
Happy to share my next video, where I give a quick tour of @Doubleword_, a platform designed for running open-source models with both Playground and API-based batch scheduling options.
🎥 Full walkthrough in the video - youtu.be/JsdJjSqhVCI
We’re pushing efficiency in large-scale LLM serving.
Our new work, QueueSpec, drafts tokens while requests queue to speed up decoding.
Results:
• Up to 4× speedup
• ~2× avg vs SGLang n-gram
Blog: blog.doubleword.ai/queue-speculat…
Code: github.com/jamesdborin/sg…
We’re focused on making large-scale LLM inference cheaper.
Our latest research, ZeroDP, boosts throughput by up to 70% by offloading weights to neighboring GPUs and fetching them just-in-time over NVLink - no model changes required.
Blog: lnkd.in/e8cGhkyE
We ran an experiment: semantic search over 2.4M arXiv papers using LLM judgments instead of embeddings.
Batch inference makes it cheap enough to ask models “is this relevant?” at query time.
⏱ Minutes
📷 < $0.01
Write-up: blog.doubleword.ai/arxiv-llm-sear…
Today we reduced the cost of our most intelligent model (Qwen3-235*) from $0.2/M - $0.6/M to $0.1/M - $0.4/M.
As we make our stack more efficient, we pass on those savings to our customers.
You can try it out at app.doubleword.ai
How many users can my GPU really serve?
In episode 2, Chief Scientist Jamie Dborin focuses on how many users you can realistically support at different context lengths and the different techniques to increase this capacity.
Watch the full episode: youtu.be/cDexxIHsUl4
What is it important to be observing when productionising AI models?
In the latest episode, Chief Scientist Jamie Dborin focuses on observability and which key signals teams working with AI in production should be looking at.
Episode 1: youtu.be/3W7VNHTYHp8?fe…#GenAI
Will you be at #SnowflakeSummit? Stop by Booth #2407 June 2 - 5 in the Expo Hall to chat with our team and learn more about self-hosted AI inference!
Please join us at booth #2407 to learn more about how we're working with Snowflake to support enterprise customers.
The BarryByte podcast returns‼️
Delighted to have advice for budding founders in the Vale from one of the UK's superstar AI founders, @MeryemArik9, fresh from a major investment in UK AI trailblazer @Doubleword_⚡📈
🎧 Listen to the full episode on Spotify:
open.spotify.com/show/5fpBHucDk…
273 Followers 917 FollowingFounder/CEO at Cleria. Ex @uber , @CocaCola, @cambridge_Uni , @lancasteruni . Interested in AI , consumer research & data science.
2K Followers 3K FollowingInvesting in advanced technologies for people and planet @kintsugiad 🏞️ Previously Co-founder of PredictionIO 🐸 (Acquired by Salesforce) and Ph.D. @UCLCS 🖥
197 Followers 1K FollowingBuilding Epsilab | ex-SWE & ML at Tower Research Capital & XY Sense | @startmate W21 Fellow | @UNSW & @ourANU CS Grad | Ultra Runner
1K Followers 5K FollowingPartner at @TrueSightVC. Active Angel. Investor in: @wearebondaval, @deel, @joinodf, @artificiallabs, @kota_benefits and more 🚀
80 Followers 217 FollowingBuilding what the scaling era missed.
Different architecture. Different bet.
Founder Mumbrane
Deploying Multi Domain Digital Employees
12K Followers 5K FollowingA digital and print magazine helping tech startups connect the dots on their entrepreneurial journeys 👩🚀👨🚀⚡️ Our Awards - @TheHustleAwards
723 Followers 713 FollowingIntel Ignite is no longer active. For startup related inquiries, please contact [email protected]. In Israel please contact [email protected]
1K Followers 765 FollowingYour global support network of AI practitioners - from engineers to business leadership - paving the way for AI in industry.
https://t.co/42yQTyLvh5
515K Followers 2K FollowingAccenture's only official global account on X. We’re not active here, but the journey continues elsewhere! Follow us on LinkedIn for our news and insights.
17K Followers 2K FollowingLondon-based accelerator uniting FinTechs with big companies | Run by @Accenture
Applications open now! https://t.co/erAagXXys6
2K Followers 52 FollowingOpen-source AI orchestration framework by @deepset_ai.
Build context-engineered agents & RAG systems in Python.
Discord for support → https://t.co/19wuHcilYP