DuckRouter

Inspiration# 🧠 AI Career Advisor - DuckRouter

Team YOGS

👥 Team Members

• Sneha Dharne – sdharne@stevens.edu
  
• Gunik Luthra – gluthra@stevens.edu
  
• Yash Gandhi – ygandhi3@stevens.edu
  
• Om Gandhi – ogandhi@stevens.edu
  

---

📌 Problem Statement

Navigating careers is tough. Our goal is to build an AI-powered system that can answer complex career questions, assess earning potential, and visualize growth trajectories—all through natural, conversational interactions.

---

💡 Project Description

We’ve developed DuckRouter, a multi-agent AI chatbot designed to give personalized career advice. It uses ChatGPT-based agents enhanced with real-world data from BLS, Levels.fyi, and LinkedIn to generate meaningful insights.

✅ The chatbot accepts inputs like:

• Profession
  
• Years of Experience
  
• Current Salary
  
• Location
  

Based on this, users can ask:

• “Should I switch careers?”
  
• “Where will I be paid more based on location?”
  
• “What’s my optimal path to earning more?”
  ...and much more.

---

🛠️ Technologies Used

• AI Agents (LangGraph + ChatGPT)
  
• Python
  
• Markov Decision Models
  
• FastAPI
  
• React.js

---

🧩 Architecture Overview

DuckRouter is modular by design, featuring a multi-agent architecture to ensure scalability and intelligent decision-making.

🔹 1️⃣ UI Interface Agent

Acts as the central coordinator that:

• Interprets user intent
  
• Gathers missing inputs via follow-ups
  
• Routes queries to appropriate backend agents
  
• Returns final insights in a clean, actionable format
  

---

🔹 2️⃣ Career Planning Agent

Delivers:

• Personalized short-term and long-term career plans
  
• Visual pitch decks outlining career progression
  

---

🔹 3️⃣ Upskilling Recommendation Agent

Bridges skill gaps by:

• Recommending courses and certifications
  
• Recommending skills to achieve the goal

---

🔹 4️⃣ Comparative Salary Agent

Provides accurate benchmarking using:

• BLS statistics
  
• Levels.fyi compensation data
  
• LinkedIn job insights
  

Factors considered:

• Role seniority
  
• Geographic location
  
• Job complexity
  

---

🔹 5️⃣ Career Switch Simulator Agent

Predicts career transitions using a Markov Model, including:

• Transition probabilities across roles/industries
  
• Real-world career switching trends
  
• Simulated outcomes for informed decision-making
  

---

🔁 Flow Summary

1. User initiates a conversation via the UI Agent
   
2. UI Agent interprets and collects missing inputs
   
3. Query is routed to the relevant backend agent
   
4. Backend agent processes the request and pulls data
   
5. UI Agent returns results in an intuitive format
   

---

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for DuckRouter

Built With

• ai
• aiagents
• fastapi
• javascript
• langgraph
• markov
• python
• react