AI Scheduling Assistant
Autonomous scheduling agent using DeepSeek 7B and Meta LLaMA 3.1 for cross-timezone meeting coordination.
Project Overview
Developed an intelligent scheduling system that tackles the challenge of cross-timezone meeting coordination by integrating advanced LLMs for natural language understanding and automated conflict resolution.
The system leverages DeepSeek 7B and Meta LLaMA 3.1 models to understand scheduling requests in natural language, automatically detect conflicts, and propose optimal meeting times across different time zones.
Deployed on AMD MI300 GPUs with vLLM inference servers, achieving sub-2 second median latency under concurrent multi-user load, making it suitable for real-time enterprise applications.
Key Features
- Natural language processing for meeting requests using advanced LLMs
- Automatic conflict detection and resolution via Google Calendar API integration
- Cross-timezone coordination with intelligent time slot suggestions
- Real-time inference with sub-2 second response times
- Prompt sanitization and response verification for safe outputs
- Comprehensive documentation and open-source reproducibility
Technologies Used
Project Details
Client
Personal Project
Timeline
1 month (Jul 2025)
Role
AI/ML Developer
More Projects
© 2025 Jane Doe. All rights reserved.