Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Model Gateway: Model Management Platform for Model-Driven Drug Discovery

Yan-Shiun Wu, Nathan A. Morin

TL;DR

The paper presents Model Gateway, a cloud-based model-management platform tailored for drug discovery that integrates LLM Agents and Generative AI tools into MLOps pipelines. It details a Kubernetes-based, containerized ecosystem with modules for versioning, access control, metadata, asynchronous execution, and dynamic consensus, plus admin and owner interfaces. An experimental evaluation demonstrates the platform handling up to 10K asynchronous submissions with 100% success and highlights scalability limits as concurrency grows, suggesting vertical and horizontal scaling as avenues for expansion. The work argues that centralized, governance-focused model management can accelerate drug discovery by improving reproducibility, accessibility, and integration across diverse tools and workflows.

Abstract

This paper presents the Model Gateway, a management platform for managing machine learning (ML) and scientific computational models in the drug discovery pipeline. The platform supports Large Language Model (LLM) Agents and Generative AI-based tools to perform ML model management tasks in our Machine Learning operations (MLOps) pipelines, such as the dynamic consensus model, a model that aggregates several scientific computational models, registration and management, retrieving model information, asynchronous submission/execution of models, and receiving results once the model complete executions. The platform includes a Model Owner Control Panel, Platform Admin Tools, and Model Gateway API service for interacting with the platform and tracking model execution. The platform achieves a 0% failure rate when testing scaling beyond 10k simultaneous application clients consume models. The Model Gateway is a fundamental part of our model-driven drug discovery pipeline. It has the potential to significantly accelerate the development of new drugs with the maturity of our MLOps infrastructure and the integration of LLM Agents and Generative AI tools.

Model Gateway: Model Management Platform for Model-Driven Drug Discovery

TL;DR

The paper presents Model Gateway, a cloud-based model-management platform tailored for drug discovery that integrates LLM Agents and Generative AI tools into MLOps pipelines. It details a Kubernetes-based, containerized ecosystem with modules for versioning, access control, metadata, asynchronous execution, and dynamic consensus, plus admin and owner interfaces. An experimental evaluation demonstrates the platform handling up to 10K asynchronous submissions with 100% success and highlights scalability limits as concurrency grows, suggesting vertical and horizontal scaling as avenues for expansion. The work argues that centralized, governance-focused model management can accelerate drug discovery by improving reproducibility, accessibility, and integration across diverse tools and workflows.

Abstract

This paper presents the Model Gateway, a management platform for managing machine learning (ML) and scientific computational models in the drug discovery pipeline. The platform supports Large Language Model (LLM) Agents and Generative AI-based tools to perform ML model management tasks in our Machine Learning operations (MLOps) pipelines, such as the dynamic consensus model, a model that aggregates several scientific computational models, registration and management, retrieving model information, asynchronous submission/execution of models, and receiving results once the model complete executions. The platform includes a Model Owner Control Panel, Platform Admin Tools, and Model Gateway API service for interacting with the platform and tracking model execution. The platform achieves a 0% failure rate when testing scaling beyond 10k simultaneous application clients consume models. The Model Gateway is a fundamental part of our model-driven drug discovery pipeline. It has the potential to significantly accelerate the development of new drugs with the maturity of our MLOps infrastructure and the integration of LLM Agents and Generative AI tools.

Paper Structure

This paper contains 19 sections, 1 equation, 7 figures, 4 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Model Versioning
  5. Model Access Control
  6. Model Metadata
  7. Asynchronous Model Execution
  8. Model Scaling
  9. Dynamic Consensus Model Management
  10. Platform Admins Tools
  11. Model Owner Control Panel
  12. Experimental Results
  13. Experimental Environment Setup
  14. Job submission Performance
  15. Job Result Performance
  16. ...and 4 more sections

Figures (7)

  • Figure 1: Model Gateway Eco-System: (Top) LLM Agents/Generative AI Applications; (Middle) Model Gateway; (Bottom) ML/computational Models.
  • Figure 2: Asynchronous Model Execution Concept
  • Figure 3: Swagger Documentation: Simple API Testing WebUI Tool
  • Figure 4: Job Monitoring
  • Figure 5: Locust Load Testing: (Top) Total Requests/Second and Failures/Second; (Middle) Response Times; (Bottom) Numbner of Users
  • ...and 2 more figures