Acumos ML Platform

Contributed to the Acumos Machine Learning Model project as an Individual Contributor for The Linux Foundation, focusing specifically on license management and federated training software components. Acumos is an open source platform that standardizes the process of creating, sharing, and deploying AI applications across enterprise environments.

My primary contribution involved developing a comprehensive license management system that could track how AI/ML models were being used across federated networks. This required creating a complete sandbox environment of the Acumos infrastructure to test the entire pipeline - from initial Python files through dockerized Python code deployment to running models over federated connections across different companies.

The license management challenge was complex: we needed to monitor model usage in real-time, update Elasticsearch logs with detailed usage analytics, and provide comprehensive usage data required for enterprise license compliance. The solution involved implementing ODRL (Open Digital Rights Language) standards for describing model rights in a way that could be easily extended and understood across different organizational boundaries.

The federated learning component required building secure, scalable connections between different company infrastructures while maintaining detailed audit trails of model usage, performance metrics, and licensing compliance. This work laid the foundation for enterprise AI/ML model sharing with proper rights management and usage tracking.