International Journal of Scientific Research and Engineering Development

The recent rise of large language models (LLMs) and autonomous multi-agent frameworks has transformed software development and project automation. This paper presents the design and implementation of the Agentic AI Project Manager, an autonomous orchestration system that can generate, assign, execute, test, and deliver full-stack web and AI/ML projects end to-end using locally hosted language models. The proposed system integrates multiple open-source 7B models such as Mistral, StarCoder2, and CodeLlama with CrewAI orchestration for multi-role task delegation. Each agent assumes a specific role, ranging from project planning and development to integration, testing, and delivery. The framework automatically generates complete functional projects, including web applications, backend APIs, machine learning pipelines, and documentation, all delivered as a downloadable ZIP file. The system is designed for local execution with configurable timeout settings to accommodate limited hardware resources. This work demonstrates a step toward practical, resource-aware AI-driven project automation capable of operating offline, ensuring privacy and reproducibility.

Aryan Kamble,"Agentic AI Project Manager: Multi-Agent Orchestration with Local LLMs" International Journal of Scientific Research and Engineering Development, V9(1): Page(1108-1112) Jan-Feb 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.