←
Azeez377 posts
Azeez avatar

Azeez

@AtlasInference

Building Atlas, pure Rust and CUDA/ROCm kernels on DGX Spark & Strix Halo | MLPerf Agentic Edge Taskforce | Ambassador @Alibaba_Qwen | Ambassador @DevinAI

◷ Joined March 2026▧ 56 Photos and videos
Jason McCartney avatar
⟳ Azeez retweeted
Jason McCartney@jmacftw·Nov 13

All of our SparkScore demos are real-time, recorded against a serving endpoint on one of our @NVIDIAAI DGX Sparks. SparkScore technology has been contributed to @AtlasInference - the O.G. of Rust-based inference engines. github.com/enntity/sparkscor…

Azeez avatar
Azeez@AtlasInference·Nov 10

Give early-access Qwen4 an @alibaba_cloud models crew like Image, Wan-2.7, Audio, Omni and see 🎥 It storyboarded, shot 10 clips, scored/re-scored after hearing a first mix, caught and fixed its own broken first cut, and signed off 🫡 @Alibaba_Qwen is cooking... can't wait 🍿

0:30
Jason McCartney avatar
⟳ Azeez retweeted
Jason McCartney@jmacftw·Nov 10

We started an open effort to train a DFlash2 drafter for GLM 5.3 Flash. It's already 92-95% as fast as one of the best drafters for the model and totally open licensed. Plug and play with most model engines and works great with our SparkGLM and @AtlasInference .

Post attachment

Enntity Lab experiment: can two @NVIDIAAI DGX Sparks train a real speculative-decoding drafter for one of the top performing local coding models from scratch? Three weeks - and about a dozen small training sessions later - our GLM 5.3 Flash drafter runs at 92-95% of the best public drafter's speed, and is already faster on coding-agent turns. Open code, open distributed training (if you want to get in), Apache 2.0 weights 🧵

Jason McCartney avatar
⟳ Azeez retweeted
Jason McCartney@jmacftw·Nov 9

We have a 27B decision model reading a 12,000-token handbook on every request and making a decision in 191ms on one @NVIDIAAI DGX Spark. It follows the rules every time and the model hasn't been fine-tuned. Here's what we did. 🧵

Jason McCartney avatar
⟳ Azeez retweeted
Jason McCartney@jmacftw·Nov 7

“I wanna go fast!” - Ricky Bobby We got a very smart 27B decision model answering in 146 ms. On one DGX Spark. On your desk. The video shows real time serving of an actual endpoint. The possibilities are endless! @AtlasInference

1:32
Azeez avatar
Azeez@AtlasInference·Nov 5

SWE-2 is already fantastic, so excited to blaze through kernel dev with this flash variant 🔥

Introducing SWE-2 Flash, our new fast & efficient coding model. On leading evals, it outperforms all American open-weight models, even ones 3x larger. To balance transparency and cyber risks, we are building a program to securely share the weights with partners & researchers.

Post attachment
Azeez avatar
Azeez@AtlasInference·Nov 4

FP4-Local=better quality/BFCL accuracy, smaller footprint 🤏 TP=2, @Alibaba_Qwen's 3.8-Flash-Next 108 tok/s code, 260+ for 4 users 📈 Prefill ~2.3× vLLM, one 17K prompt, 5.8 vs 9.2s 💻 Worked w/ @jmacftw, >NVFP4 🧠 Try it: sparkrun run @atlas/qwen3.8-flash-next-fp4-local-tp2

HackerHouse Campus avatar
⟳ Azeez retweeted
HackerHouse Campus@HackerHouseCo·Oct 27

we brought @cognition to @GeorgiaTech 🚀 100+ students teamed up with @DevinAI for a 7 hour hackathon at @HackerHouseCo and shipped 25 projects. everyone left with devin credits and a pro plan for the month. - 1st place: AKA studio, describe a ring and get real CAD files back. akhilesh, armaan, greg, @Keylemx - 2nd place: Asymmetry, personal ai agents that connect your needs to your network - 3rd place: Raven, real-time emergency detection from traffic feeds and public alerts thank you @dabit3, azeez @atlasInference and the cognition team for working with us. :)

0:27
Azeez avatar
Azeez@AtlasInference·Oct 27

"I no longer think we're going to be able to just patch all the bugs." - Sam Altman We love @vllm_project, but 40 advisories in 6 weeks. 5 rated high, 3 let an attacker run code on YOUR server. Atlas is just Rust + CUDA/ROCm kernels, no dependencies. None of the 40 touch it :)

Post attachment
Azeez avatar
Azeez@AtlasInference·Oct 23

@RisingSayak would love to look into integrating this for GB10 as fused PiGemma (AdaRMS + gated residual/flow-matching step) so @LeRobotHF users get native kernelize() speedups out of the box!

Azeez avatar
Azeez@AtlasInference·Oct 23

Attention robotics fans 🚨 VLA's like @physical_int Pi05 are coming to Atlas 🤖 Watch Atlas control its first robot on GB10 solely through the model itself and the vision feeds ⚙️ DGX Sparks have many use cases beyond text! AGX Thors aren't the only physical edge devices 🫢

0:04
Azeez avatar
Azeez@AtlasInference·Oct 21

It's amazing to see +1 Rust-backed engine takeover :) Early 2026 Atlas came about as a pure CUDA zero dependency framework. I'd say we're worth a (re)try too. Just 30s to first token 😉 GLM5.3, Qwen3*, Nemotron3* github.com/Atlas-Inf/atlas sparkrun run @atlas/qwen3.8-27b-nvfp4

🚨 @NVIDIAAI just quietly dropped its own local inference server... No announcement. Single-digit stars It's called Flare and Apache 2.0 Built for SM120/SM121 GPUs RTX PRO 6000 Blackwell and DGX Spark >PyTorch + FlashInfer at build time >Rust runtime at serve time. No Python. >CUDA graphs captured before serving >FP8 KV cache by default >Recipes for Qwen Flash, GLM and DeepSeek github.com/NVIDIA/flare-loca…