QwerkyAI

Almost 1 month out from the big race! Get into the derby mode, head over to llm-derby.web.app Built by: @QWERKYAI What is it: it's an AI model horse race app where LLMs compete head-to-head on MMLU-Pro Hard benchmarks and you can bet virtual tokens on your favorite model 🏇 While you watch the race double check make sure your Rolex is real at jolly-lolly-f1d0b3.netlify.app Powered by: @QWERKYAI

Thinking about your next NVIDIA Certification in 2026? Join our certification experts for a global webinar on April 30 to get an overview of refreshed and upcoming certifications in data science, physical AI, OpenUSD, and AI infrastructure. You’ll also get prep resources and practical guidance to help you succeed, including sample questions, study tips, and a live Q&A with the experts. 👉 Save your seat and start planning your 2026 certification goals. 🔗 nvda.ws/4sqT2VJ

Here's to Qwerky AI at @nvidia GTC 2026 in the books! 🍻 Our team met incredible people, had great conversations, and is leaving San Jose excited to bring the next phase of QWERKY AI to reality. A massive thanks to our Inference Happy Hour co-hosts for making it the highlight of the week: @MassedCompute, @rapt_a, Amaara Networks Inc., and @Modular. 🙌 And thank you to our @QWERKYAI GTC 2026 team! You all certainly got your steps in.🚶‍♂️@ulmentflam, @jandrewstrauss, Derek Bronson, and Kaitlin McKnight. If we didn't get a chance to connect this week, our DMs are open. We'd love to keep the conversation going. #GTC2026 #NVIDIA #NVIDIAInception #AI

We're co-hosting an Inference Happy Hour at GTC tonight! Come hang, grab a drink, and talk AI with the QWERKY AI team. We'll be at the Farmers Union until 9 PM PDT: luma.com/drfpq7kh! Check out more about QWERKY AI here: qwerky.ai/gtc-2026 #GTC2026 #NVIDIA #NVIDIAInception #GenerativeAI

NVIDIA GTC 2026 is in full swing! 🤙 @QWERKYAI's CTO, @ulmentflam, has been soaking up sessions all week, starting with OpenModels in Health Care. If you're here in San Jose this week, he'd love to show you what we've been building. @QWERKYAI is making models 5x faster and 80% cheaper. Schedule some time before GTC is over to meet up with Evan for a demo of QRe 👉 qwerky.ai/gtc-2026 #GTC2026 #NVIDIA #NVIDIAInception #AI

The @QWERKYAI team has been busy busy since we landed in San Jose for @NVIDIAGTC 2026, but we wanted to share some team pics from the week so far! 🙌 We'd love to meet up and demo you our tech, making LLMs run 5x faster and 80% cheaper 👉 qwerky.ai/gtc-2026 #NVIDIAGTC #GTC2026 #NVIDIA #NVIDIAInception

We're heading to @NVIDIAGTC next week! Qwerky AI is co-hosting the Inference Happy Hour alongside Massed Compute, Modular, Rapt AI, and Amaara Networks on Wednesday March 18th at Farmers Union in San Jose. If you're building at the inference layer, whether that's optimization, deployment, or pushing the boundaries of what's possible per dollar per token, this is your room. RSVP here: luma.com/drfpq7kh See you in San Jose. #Modular #MassedCompute #RaptAI #AmaaraNetworks #AI #NVIDIAGTC

MAX was originally architected around transformer-based models. @QWERKYAI needed state space model support, so they built it: eight custom kernels in two weeks. 😮 Dig into their learnings from establishing first-class SSM support in MAX: qwerky.ai/blog/bringing-…

AI's next chapter is being written in the Columbia area. With support from partners like the Boyd Innovation Ctr to funding from SC Launch, novel ideas become scalable ventures that change the world. Check out a recent feature on local @QWERKYAI, 🤖 postandcourier.com/columbia/busin…

Join us today at 9:30 AM PT in the Modular forum for an Ask Us Anything session with @clattner_llvm and @chaoyu_ about our recent acquisition of @bentomlai! We'll answer your questions live and share our vision for the future. 🔮 forum.modular.com/t/modular-has-…

🚀 BentoML is joining @Modular. Together, we’re building a complete stack for production AI inference. Fast, portable, and free from hardware lock-in. Learn more 👉 bentoml.com/blog/bentoml-i…

@QWERKYAI is so thrilled to join the impressive ranks of SC Launch Inc.’s portfolio and contribute to South Carolina innovation! #SCInnovation #SustainableAI #HumanCenteredAI #TechStartup #AIStartup #SouthCarolinaTech #MadeInSC #AIInvestment

We just shipped the first CPU-only Selective Scan kernels for state space models, built Mamba support into @Modular 's MAX framework in two weeks, and validated eight custom kernels against @tri_dao's reference implementation. First write-up on our implementation from our CTO, @ulmentflam, on the blog here: qwerky.ai/blog/bringing-… #SSM #StateSpaceModels #GenAI #AI #LLM #LargeLanguageModels #MAX

@QWERKYAI teams up with Inbox Beverage to power an AI-driven design and customer engagement platform using a custom 8B-parameter model for smarter craft beverage planning. 🔗 tinyurl.com/2s3d2tsc #AI #Partnership #CraftBeverage #CustomerEngagement #Innovation

Qwerky AI teams up with @InboxBeverage to build an AI-powered platform for custom turnkey containers! Our 8B param state-space model delivers faster, more efficient results for instant layouts, troubleshooting, and ops wizardry. Slashing design times, boosting ROI in a cutthroat market. MaxFann CEO: “This partnership is about giving operators better answers, faster.” inboxbeverage.com prnewswire.com/news-releases/…

Modular 26.1 is a game-changer. We’re proud and excited to see @QWERKYAI's work featured and to be part of this innovative community! 🔥We're bringing Mamba architecture and state-space models to MAX—exploring alternatives to traditional Transformers with faster inference and much longer context windows. Our team developed custom Mojo kernels, new model architectures, and serving pipeline enhancements. Check out more about @Modular 26.1 here: modular.com/blog/26-1-rele… #AI #MachineLearning #mamba #StateSpaceModels #Mojo

Debug in eager mode. Compile with one line. Deploy to NVIDIA, AMD, or Apple silicon. MAX 26.1 launches today, delivering all three 🎯 The Python API is now stable. PyTorch-like modeling that compiles to production performance across GPU vendors. Community contributors are extending it with Qwen3, BERT, Mamba, and visual generation pipelines. We've open-sourced everything: the API, GPU kernel library, model architectures, and serving pipelines. More about this release ⤵️ modular.com/blog/26-1-rele…

Album Release Party Coming Soon! Stay Tuned #QwerkyAI #ComputeDrop #AI #Optimization #VRAM