Pluralis Research works on Protocol Learning: training and serving large models in a fully decentralized way on small consumer-grade devices connected via the internet. Despite being dismissed as infeasible, we have made significant advances on this problem, most recently Agora, a permissionless run that pretrained an 8B model from scratch on consumer GPUs spread over the internet, with no single participant ever holding the full weights (tech report). While many of the core research problems have been solved, Protocol Learning unlocks a series of new challenges. For the mission in full, read A Third Path: Protocol Learning.

Our training and inference doesn't happen in a datacenter. It happens on consumer nodes and cloud instances that are not co-located, connected by ordinary internet, joining and leaving mid-run. Your primary role is to architect, build, and scale the platform that keeps continuous experimentation and large-scale training running on top of that: infrastructure orchestration, distributed compute, and the services that tie them together.

Key Responsibilities

• Multi-cloud infrastructure: Design the resource management systems that provision and orchestrate compute across AWS, GCP, and Azure with infrastructure-as-code (Pulumi/Terraform). Handle dynamic scaling, state synchronization, and concurrent operations across hundreds of heterogeneous nodes.
• Distributed training and inference systems: Architect fault-tolerant infrastructure for distributed ML. GPU clusters, NVIDIA runtime, S3 checkpointing, large-dataset management and streaming, health monitoring, and resilient retry strategies.
• Real-world networking: Build the systems that simulate and handle real network conditions such as bandwidth shaping, latency injection, packet loss. Managing node churn and keeping data flowing across workers with heterogeneous connectivity.

What We're Looking For

• Infrastructure and platform engineering (required): Production experience with infrastructure-as-code (Pulumi/Terraform/CloudFormation) managing multi-cloud deployments, Docker/Kubernetes (EKS), GPU workloads, and heterogeneous clusters at scale.
• Distributed systems and ML infrastructure: You understand distributed training workflows: checkpointing, data sharding, model versioning, long-running job orchestration.