AI Researcher Take-Home Assessment

Project Overview

This project demonstrates a lightweight multi-agent AI workflow for autonomous logistics rerouting using LangGraph.

The system automatically responds to shipment delays by analyzing telemetry data, evaluating alternative carriers, and selecting the best rerouting option.

---

Business Problem

Logistics companies often experience shipment delays due to weather conditions, carrier failures, or operational issues.

Instead of relying on manual intervention, this proof-of-concept demonstrates how AI agents can autonomously:

• Detect shipment delays
• Analyze the disruption
• Evaluate alternative carriers
• Select the optimal carrier
• Execute the rerouting process
• Notify the logistics system

---

Technology Stack

• Python 3.11+
• LangGraph
• LangChain
• OpenAI (Optional)
• python-dotenv

---

Project Structure

AI-Researcher-Assessment/
│
├── app.py
├── agents.py
├── workflow.py
├── config.py
├── requirements.txt
├── README.md
├── .env.example
│
├── report/
│   └── Research_Report.md
│
└── presentation/

---

AI Workflow

Delay Alert
      │
      ▼
Telemetry Agent
      │
      ▼
Decision Agent
      │
      ▼
Execution Agent
      │
      ▼
Shipment Successfully Rerouted

---

Agent Responsibilities

1. Telemetry Agent

• Detect shipment delay
• Read shipment information
• Determine delay severity

2. Decision Agent

• Evaluate available carriers
• Compare ETA
• Compare cost
• Compare reliability
• Select the best carrier

3. Execution Agent

• Execute rerouting
• Update shipment
• Generate confirmation

---

Installation

Install dependencies

pip install -r requirements.txt

---

Configuration

Copy

.env.example

to

.env

If you have an OpenAI API Key

OPENAI_API_KEY=your_key
USE_MOCK=false

Otherwise keep

USE_MOCK=true

The project works without any API keys.

---

Run

python app.py

---

Expected Output

The workflow will

• Detect shipment delay
• Evaluate carrier options
• Select the best carrier
• Execute rerouting
• Display workflow summary

---

AI Tool Usage Decision Log

AI tools were used only to accelerate development and documentation.

• ChatGPT was used for brainstorming architecture, improving code quality, generating documentation, and refining the research report.
• Claude was used to assist in structuring the presentation deck.
• Final workflow design, evaluation methodology, architectural decisions, and implementation were reviewed and finalized manually.

---

Future Improvements

• Real-time carrier APIs
• Database integration
• Human approval workflow
• Retrieval-Augmented Generation (RAG)
• Multi-modal telemetry inputs
• Real-time monitoring dashboard

---

Author

AI Researcher Intern Take-Home Assessment