$ cat project.txt
AI ADS RECOMMENDS
AI-powered advertisement recommendation system. Uses LLM agents and vector embeddings to surface the most relevant ads for each advertiser and context.
$ ls -la project/
A recommendation platform that analyzes advertiser data and user context to select optimal ads through multi-agent LLM orchestration.
The Challenge
Traditional ad systems lack semantic understanding of ad-user-context relationships, leading to low relevance and wasted spend.
Our Solution
A multi-agent backend evaluates ads against context, with vector embeddings for semantic retrieval and PostgreSQL persistence, served through a Nuxt chat-style interface.
Key Features
Technical Architecture
Frontend
- framework: Nuxt 4
- stateManagement: Pinia
- styling: Tailwind CSS
- testing: N/A
Backend
- runtime: Python
- framework: FastAPI + Celery
- database: PostgreSQL
- caching: Redis
Deployment
- hosting: Docker
- ci_cd: GitHub Actions
- monitoring: N/A
Results & Impact
Semantic matching improves ad relevance over keyword rules
Multi-agent supervision keeps recommendations on-context
Vector retrieval scales to large ad inventories
Containerized stack streamlines local development
Gallery
Recommendation dashboard
Conversational ad matching