-
-
ClickHouse database
Inspiration
My background is in economics — I've always been fascinated by what drives the world economy, why markets move, and how global events translate into price action. That curiosity led me to study how financial systems actually work under the hood.
Earlier this year, I attended the Morgan Stanley Technology Spring Insight Week, where I got exposure to the key software infrastructure behind real trading platforms — real-time data feeds, alert systems, and autonomous monitoring tools that traders rely on. That experience made me want to build something that bridges my love of finance with my growing technical skills.
Multiagent Market Monitor is the result of that intersection. An autonomous system that watches markets 24/7, stores historical price data in ClickHouse Cloud, enriches alerts with live news context, and fires SMS notifications the moment a significant price movement is detected — the kind of tool that sits behind every serious trading desk.
What it does
Multi-agent Market Monitor is a 5-agent autonomous pipeline that fetches live stock prices every 5 minutes, stores historical data in ClickHouse Cloud, enriches analysis with real-time news context, and sends SMS alerts when significant price movements are detected.
How we built it
Built in Python with 5 specialist agents: Fetcher (Alpha Vantage API), ClickHouse Agent (data storage), Research Agent (NewsAPI), Analyser Agent (threshold detection), and Notifier Agent (Twilio SMS). Each agent has a single responsibility and communicates through a shared data pipeline.
Challenges
Integrating ClickHouse Cloud with real-time data ingestion, managing Alpha Vantage API rate limits across multiple stocks, and wiring 5 agents into a clean autonomous loop without race conditions.
What we learned
How to architect true multi-agent systems, real-time data pipelines with ClickHouse, and the importance of autonomous agent orchestration in fintech applications.
Log in or sign up for Devpost to join the conversation.