AI Pulse
Overview
Autonomous multi-agent AI news aggregation and personalization platform.
AI Pulse is a production-ready multi-agent AI system that autonomously aggregates, ranks, summarizes, and delivers curated AI/tech news. The system orchestrates three specialized AI agents — Digest Generator, Content Curator, and Email Synthesizer — powered by Gemini 2.5 Flash. It processes articles from multiple sources including YouTube, OpenAI Blog, and Anthropic feeds, applies semantic ranking based on user profiles, and delivers personalized digests via automated SMTP workflows.
The Challenge
Tech professionals struggle to stay updated with high-quality AI content scattered across blogs, research platforms, and video channels. Manual filtering is time-consuming and inefficient.
Obstacles Faced
Designing coordination between multiple AI agents
Avoiding redundant or low-quality content aggregation
Implementing personalized semantic ranking
Building reliable scraping pipelines
Ensuring automated email delivery at scale
Feature Engine
- Architected a 3-agent pipeline: Digest Generator, Curator, and Email Synthesizer
- Implemented semantic user scoring system (0–10 relevance score)
- Built scraping infrastructure using BeautifulSoup4
- Used PostgreSQL with SQLAlchemy for structured content processing
Outcomes & Impact
Processed 100+ weekly AI articles and videos
Delivered fully automated personalized AI digests
Reduced manual news filtering effort significantly
System Architecture
Agent-based AI orchestration model
Hybrid stack (Python AI services + MERN frontend)
Content ingestion → ranking → summarization → delivery pipeline
Docker-based modular service isolation
Processed 50+ daily articles across multiple feeds
Interface Snapshots
Execution Lessons
Designing agentic AI systems with clear responsibility boundaries
Semantic ranking strategies using LLMs
Handling scraping reliability and rate limits
Building production-ready automated pipelines
Orchestrating AI workflows beyond simple prompt-response systems