‹ All Projects
Project Thread

CrewQuest: AI Tycoon — Gamified Multi-Agent Corporate Workforce Simulation

🐙 View GitHub

Overview

A real-time visual simulation of an autonomous AI workforce. Users act as a CEO, assigning complex tasks to a hierarchy of 11 specialized AI agents who collaborate to execute strategies.

Tech Stack

Python (CrewAI Framework)Streamlit (Frontend & Execution Engine)Google Gemini 2.5 FlashPython-dotenv

Features

  • Corporate Hierarchy: 11 distinct roles including Executive Secretary and Department Managers.
  • Dynamic Staffing: Real-time delegation strategy adjustment based on available staff.
  • Live Thought Animation: Sequential rendering of Thought, Tool Usage, and Action cycles.
  • Custom Callback Engine: Intercepts and parses low-level agent .log files for UI display.

Key Decisions

01

Built a custom log-parsing orchestrator because the default CrewAI Manager hid the granular communication between department managers.

02

Integrated DiceBear procedural avatars to give each of the 11 roles a unique visual identity that updates dynamically based on department selection.

Learnings

→Gained expertise in forcing LLMs to output strictly structured delegation plans through specialized "Boss Instructions" prompting.
→Solved the "Black Box" transparency issue by implementing real-time UI chat bubble rendering via custom log interceptions.
EmailGitHubLinkedInResume