Umar Mohamed Iyoob Umar

Title: Umar Mohamed Iyoob Umar and AI-Driven MLOps

Introduction

Umar Mohamed Iyoob Umar is an inventor based in Pflugerville, Texas. His patent work addresses the management of machine learning operations for cost forecasting models in multi-tenant cloud environments. He is associated with Kyndryl, Inc.

Latest Patents

Umar is named on one patent, titled “AI Customized MLOps Framework in a Multi-Tenant Cloud Environment.” The invention describes a computer-implemented system, method, and program product for managing machine learning operations for cost forecasting models. It receives a tenant’s request for a subscription machine learning model, uses the tenant’s configuration profile to select a model from a registry, and deploys that model for the tenant. The system monitors model usage at a predetermined refresh frequency and evaluates whether the model exceeds a specified accuracy threshold. When the threshold is exceeded, it automatically replaces the operating model with a second subscription model from the registry.

Career Highlights

Umar’s patent reflects work at the intersection of artificial intelligence, cloud computing, and machine learning operations. Its approach uses tenant-specific configuration and automated model selection and replacement to manage cost forecasting models in a multi-tenant setting. His patent portfolio includes 1 patent.

Collaborations

Umar’s listed coworkers include Omar Odibat and Mouleswara Reddy Chintakunta. The patent’s focus on tenant configurations, model registries, and automated deployment illustrates a collaborative, systems-oriented approach to MLOps.

Conclusion

Umar’s invention presents an automated framework for selecting, monitoring, and replacing machine learning models for cost forecasting in multi-tenant cloud environments.