Machine Learning Engineer - ML Training Platform

Architects and scales the platform that runs distributed, decentralized model training and inference across a mix of consumer and cloud nodes connected over public internet. Owns multi-cloud orchestration across major providers, fault-tolerant GPU training infrastructure including checkpointing, sharding, streaming, and retry strategies, and handling real-world constraints like bandwidth shaping, latency variation, and node churn. Requires production IaC experience, Kubernetes, distributed ML workflow knowledge, strong Python engineering, and observability or SRE practice.