Pluralis Research
Pluralis Research builds a protocol and tooling for collaborative, decentralized training and ownership of foundation models, enabling training and serving across heterogeneous, consumer-grade devices over the internet. The team demonstrated Agora, a permissionless run that pretrained an 8B model from scratch without any single participant holding the full weights. The privately held company is backed by Union Square Ventures and focuses on developer-facing infrastructure for large-scale model training and serving.
Open Positions at Pluralis Research
3 open positions
Build and own a geo-distributed inference stack that serves reinforcement-learning rollouts and future models across untrusted consumer hardware and public internet connections. This research-heavy role combines distributed systems engineering with novel inference algorithm design.
Research Engineer responsible for making RL post-training work across a decentralized network of consumer GPUs and Macs. You will build the end-to-end training stack, invent algorithms for asynchronous and partially trusted environments, and ship the first decentralized post-trained models.
Build the distributed training systems that take Protocol Learning from an 8B-model demonstration to frontier-scale pretraining across heterogeneous consumer hardware and ordinary internet connections. The role combines large-scale model training, performance optimization, fault tolerance, and monitoring in a remote-first research environment.