Case Study: Nucleai scales spatial biology model development and EC2 workloads with ClearML

Nucleai scales thousands of spatial biology model runs with ClearML

Nucleai, a company focused on AI-powered spatial biology research, needed to manage the complexity of running dozens of concurrent PyTorch-based deep learning pipelines for drug and biomarker discovery. Their challenge was to achieve full observability, reproducible experiments, and cost-efficient scaling across hundreds of GPU-backed EC2 instances.

By implementing ClearML as their end-to-end AI/ML orchestration and observability platform, Nucleai gained deep visibility into its pipelines and resource usage. The solution provided a single source of truth for experiments, enabled bottleneck identification through detailed scalar tracking, and offered cost-aware scalability via an AWS Autoscaler. As a result, Nucleai achieved a highly efficient platform with actionable metrics, preserved lineage, and controlled EC2 costs, allowing their team to iterate with confidence.

Nucleai