Find Remote MLOps & AI Infrastructure Jobs

A digital twin is a living simulation of a physical system — updated with real-time sensor data, calibrated against actual performance, and used to pr...

Managing 400 GPUs across three cloud regions sounds straightforward until a training job in us-east consumes 64 A100s for 72 hours, a client in eu-wes...

ML models are only as good as their labels. A classification model trained on annotations where 15% of the labels are wrong will learn the mistakes. O...

Cloud bills surprise companies every month. Training job costs spike unpredictably. Spot instance savings evaporate when availability drops. Reserved...

Vector search is easy at 1 million documents. At 500 million, everything breaks — index sizes exceed memory, query latency spikes, and the cost of re-...

We deploy ML models for 20 clients, and each deployment is a production system that real users depend on. When a model serving endpoint goes down at 2...

When an ML model misbehaves in production, the clock starts ticking. Customer trust erodes with every bad prediction served. Our observability platfor...

Our creative platform generates 2 million images per day for e-commerce product listings. Each request needs to match brand guidelines, maintain produ...

Fourteen ML teams deploy models independently, and nobody knows what's in production. Model versions are tracked in spreadsheets. Rollback procedures...

Model performance degrades silently. A retrained model scores well on validation but produces subtly worse predictions in production because the data...