This is a submission for the Hermes Agent Challenge: Build With Hermes Agent
What I Built
I built Alpha-Dairy Quant Pipeline. It is an automated system for tracking and trading dairy products (like Cheddar Blocks, Butter, and Dry Whey) using math and data.
The system automatically pulls market numbers, tests a smart trading strategy, figures out the best way to divide money between the products, and shares updates with team members. This solves the problem of trying to guess when to buy or sell food commodities by letting an AI agent safely handle the data work.
Demo
Instead of a video, here are the direct outputs and screenshots from our live project setup.
1. Automated Chat Outputs
Our system connects directly to a Discord channel. When asked to run the engine, the agent plans out the task and outputs a clean market report:
The agent creates and attaches a high-resolution chart directly to the chat to show performance against the baseline market indexes.
The agent also exports and uploads a clean spreadsheet tracking file for easy backup downloads.
Code
You can view the full repository here:
👉 GitHub Link: alpha-dairy-pipeline
My Tech Stack
AI Agent Engine: Hermes Agent Framework
Language: Python 3.12
Data Math libraries: Pandas, NumPy, and SciPy
Graphs & Visuals: Matplotlib
Data Retrieval: yFinance API
Database Store: SQLite
Platform Chat Bridge:* Discord Bot Integration
How I Used Hermes Agent
Hermes Agent acts as the intelligent director at the center of the entire project. I relied heavily on its unique agentic skills to perform real, complex work:
Multi-Step Planning: The agent maps out its own checklist (such as reading market data files, running the script, and checking metrics) without hardcoded steps.
Tool Use and Code Execution: The agent safely spawns a sandboxed python execution runtime to process thousands of data points across 28 years of history, run math optimizations, and render image files.
Autonomous Documentation: After finding that Dry Whey had massive volatility (~103% annualized), the agent automatically went in and updated our project's persistent file (MEMORY.md) with a written explanation of the statistical risk to keep the team aligned.
Top comments (0)