background
Research
Deliveringhigh-qualityAIfor8billionuniquelives.
Ourresearchisfocusedoncompilers,sharding,peer-to-peernetworks,inference,anddatageneration.
Oct 23, 2025
How Far Can One GPU Go? Running Qwen2.5-32B, Mistral 7B, and Llama-2-13B on a Single RTX 3090 How Far Can One GPU Go? Running Qwen2.5-32B, Mistral 7B, and Llama-2-13B on a Single RTX 3090
How Far Can One GPU Go? Running Qwen2.5-32B, Mistral 7B, and Llama-2-13B on a Single RTX 3090
Oct 10, 2025
mem-agent: Equipping LLM Agents with Memory Using RL mem-agent: Equipping LLM Agents with Memory Using RL
mem-agent: Equipping LLM Agents with Memory Using RL
Aug 04, 2025
Towards Open Evolutionary Agents Towards Open Evolutionary Agents
Towards Open Evolutionary Agents
May 30, 2025
Towards Feasible, Private, Distributed LLM Inference Towards Feasible, Private, Distributed LLM Inference
Towards Feasible, Private, Distributed LLM Inference
May 22, 2025
Guided Generation: Decentralized Synthetic Data Generation Guided Generation: Decentralized Synthetic Data Generation
Guided Generation: Decentralized Synthetic Data Generation
Release
May 07, 2025
DNET: Towards Heterogeneous Multiprocessing for Deep Learning DNET: Towards Heterogeneous Multiprocessing for Deep Learning
DNET: Towards Heterogeneous Multiprocessing for Deep Learning
Join the community