Update README.md

README.md

@@ -1,110 +1,171 @@
 ---
-title: AI Agent Systems Competition
-emoji: 🏢
+title: Do You Want to Build AI Agents That Run Real Businesses?
+emoji: 🚀
 colorFrom: blue
-colorTo: green
+colorTo: purple
 sdk: gradio
 pinned: true
-sdk_version: 6.5.1
 ---
 
-# AI Agent Systems Competition (Founding Team Track)
+# AI Agent Execution Competition — Participation Guide
 
-We are running a builder-focused competition to identify exceptional engineers who can design and ship **reliable AI agent systems**.  
-Top performers will be invited to join our founding team / execution center.
+**Duration:** 3-Week Build Phase + 2-Week Review Phase  
+**Location:** Virtual  
+**Host:** The Chinese University of Hong Kong, Shenzhen
 
 ---
 
-## 🎯 What to build
+## Overview
 
-Create an **AI agent system** that solves a real business execution workflow.
+The AI Agent Execution Competition is a hands-on systems challenge for builders who want to design and ship **AI agents that execute real business workflows**.
 
-Choose one track (or propose your own):
+This competition emphasizes execution, reliability, and ownership — not flashy prototypes.
 
-### Track A — Sales Follow-Through Agent
-An agent that turns inbound/outbound leads into scheduled calls and updated CRM states (with follow-ups, retries, and escalation).
+Participants will design and demonstrate AI agent systems that solve operational problems such as sales follow-through, recruiting coordination, or other measurable execution workflows.
+
+There are no monetary prizes.
+
+**Outstanding participants may be invited to join our AI systems team and collaborate on real-world agent infrastructure projects.**
+
+This competition is about identifying builders who can turn AI into operating systems for real businesses.
+
+---
+
+## Competition Tracks
+
+Participants may choose one track:
+
+### Track A — Sales Execution Agent
+
+Build an agent that manages sales workflows:
+
+- lead follow-through
+- communication coordination
+- scheduling
+- pipeline tracking
+- escalation handling
+
+Goal: maximize execution reliability and workflow completion.
+
+---
 
 ### Track B — Recruiting Coordination Agent
-An agent that coordinates candidates + interviewers across scheduling, reminders, status updates, and handoff to humans when needed.
 
-### Track C — Open Track (Execution Agent)
-Any execution workflow that has clear ownership + measurable outcomes (support escalation, onboarding, ops tracking, etc.).
+Build an agent that coordinates hiring workflows:
+
+- interview scheduling
+- candidate communication
+- reminders and follow-ups
+- status tracking
+- failure recovery
+
+Goal: reduce operational friction and improve workflow efficiency.
 
 ---
 
-## ✅ Submission requirements (must have)
+### Track C — Open Execution Agent
 
-Submit **one** entry including:
+Build an agent that solves any measurable execution workflow:
 
-1) **Code repo (required)**  
-- GitHub or Hugging Face repo link  
-- Must include: `README`, setup steps, and system architecture description
+- customer support coordination
+- onboarding automation
+- operations tracking
+- internal workflow management
 
-2) **Demo video (required, 3–5 minutes)**  
-Your video must show:
-- A live run of the agent/system
-- A walkthrough of system components
-- At least **one failure case** and how your system handles it
+Participants must clearly define the problem and success metrics.
 
-3) **Short write-up (required, 1–2 pages OR README section)**  
-Include:
-- Problem definition & assumptions  
-- Architecture diagram (can be simple)  
-- Reliability & failure handling strategy  
-- Metrics you track (and why)
+---
+
+## Schedule
+
+**Competition Timeline**
+
+Week 1–3 — Build Phase  
+Week 4–5 — Review Phase  
+End of Week 5 — Winners Announced
+
+Exact calendar dates will be posted in Discussions.
+
+---
+
+## Submission Requirements
+
+Each participant must submit:
+
+### 1. System Report (Required)
+
+A concise report (1–3 pages) describing:
+
+- problem definition
+- architecture design
+- execution workflow
+- reliability strategy
+- evaluation metrics
 
 ---
 
-## 🧪 Evaluation rubric
+### 2. Demo Video (Required, 3–5 minutes)
 
-We evaluate systems based on:
+The video must demonstrate:
 
-- **Reliability & failure handling** (timeouts, retries, fallbacks, human handoff)
-- **Execution logic & ownership** (clear workflow/state, no “prompt-only” demo)
-- **Real-world usability** (deployable mindset, clear interfaces, readable code)
-- **Outcome focus** (metrics, success criteria, measurable impact)
+- a live agent run
+- system walkthrough
+- at least one failure case and recovery behavior
 
-> Flashy UI is not required. We care about systems that can run.
+---
+
+## Evaluation Criteria
+
+Submissions are evaluated on:
+
+- reliability and failure handling
+- execution logic and ownership modeling
+- real-world usability
+- outcome focus and measurable impact
+- clarity of system design
+
+This competition rewards systems that **operate reliably**, not just impressive demos.
 
 ---
 
-## 🗓️ Timeline
+## Participation Rules
 
-- Registration deadline: **[YYYY-MM-DD]**
-- Submission deadline: **[YYYY-MM-DD]**
-- Final demos / interviews: **[YYYY-MM-DD]**
-- Winner announcements: **[YYYY-MM-DD]**
+- Individual participation only
+- Participants may use AI tools during development
+- Submissions must be original work
+- External libraries and frameworks are allowed
+- GPU resources are optional and not required
 
-(All times: Pacific Time)
+Joining this Hugging Face organization constitutes participation.
 
 ---
 
-## 📮 How to submit
+## How to Submit
 
-Fill out the submission form:
+Create a Discussion post titled:
 
-- **Submission Form:** [PASTE YOUR GOOGLE FORM LINK HERE]
+**“Competition Submission — [Your Name]”**
 
-What you will submit in the form:
-- Name / email
-- Track
-- Repo link (GitHub or HF)
-- Demo video link (YouTube / Drive / Loom)
-- Optional: short note (what you’re proud of / what’s incomplete)
+Include:
+
+- demo video link
+- PDF report link
+- short summary of your system
+
+Submissions are reviewed manually by the evaluation team.
 
 ---
 
-## 📌 Rules
+## Community & Support
 
-- Individuals or teams (up to **[N]** people)
-- You may use AI tools during development
-- Submissions must be original work (you can use open-source libraries with proper attribution)
-- We may showcase your demo publicly (only with your permission — specify in the form)
+Use the Discussions tab for announcements, questions, and technical discussion.
 
 ---
 
-## 🙌 Contact
+## Final Note
+
+This is not a showcase competition for prototypes.
 
-Questions? Post in **Discussions** or email: **[your-email]**
+It is a search for builders who can design AI systems that execute real workflows.
 
-Good luck — we’re excited to see what you build.
+We are excited to see what you build.