Pluralis Research
Machine Learning Engineer - ML Training Platform
San Francisco +1 · Remote · Full-time · Posted Aug 31, 2026
- Location
- San Francisco
Australia - Experience
- No specific number of years stated; production experience is required.
- Visa sponsorship
- Full visa sponsorship and relocation support to either Australia or the US are optional.
About the job
Architect, build, and scale the infrastructure platform for continuous experimentation and large-scale decentralized model training and inference across consumer devices and cloud instances.
What you'll do
- Design multi-cloud resource management and orchestration across AWS, GCP, and Azure.
- Build fault-tolerant distributed training and inference infrastructure, including GPU workloads, checkpointing, data management, monitoring, and retries.
- Develop networking systems for real-world conditions, including bandwidth shaping, latency, packet loss, and node churn.
What they ask for
- Production experience with infrastructure-as-code (Pulumi/Terraform/CloudFormation), multi-cloud deployments, Docker/Kubernetes (EKS), GPU workloads, and heterogeneous clusters at scale.
- Understanding of distributed training workflows, including checkpointing, data sharding, model versioning, and long-running job orchestration.
- Experience with decentralized networking concepts such as P2P, NAT traversal, traffic shaping, and real bandwidth constraints.
- Strong Python engineering, observability and SRE practice, performance profiling, and incident response.
Nice to have
- Experience with foundation model pre-training, post-training, or RL.
- Experience at proprietary, open-weight, or open-source AI labs.
Pay
High base salary (amount not stated); equity-heavy package (amount not stated)
Tools
AWS, GCP, Azure, Pulumi, Terraform, CloudFormation, Docker, Kubernetes
AI focus
Build infrastructure for decentralized training and inference of large models, including distributed ML systems and GPU workloads.
Model-written summary · the original posting is authoritative