Contact person: Business Client

Phone:Show

Email:Show

Location: Jersey City, United States

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"Automated Email-to-Quote Agent - Freight Brokerage
Overview
Build an end-to-end automated system that monitors a Gmail inbox for freight quote requests,
extracts shipment details, calculates quotes via API, generates a PDF quote document, and
replies to the email automatically. You are allowed to use any language, framework or tool that
you think can speed up the process.
Deliverable: code repository with the working implementation.
Time allocation: 2-3 hours
Scenario: Automated Freight Quote Response System
You are building an automation for a freight broker that receives quote requests via email. The
system should operate autonomously to handle incoming requests without human intervention.
Example Incoming Email
From: [login to view URL] @[login to view URL]
Subject: Freight Quote Needed - LA to Chicago
Hi,
We need a quote for shipping 2 pallets of electronics from our
warehouse in Los Angeles, CA 90021 to our distribution center
in Chicago, IL 60601.
Details:
- Total weight: 800 lbs
- Dimensions: Each pallet is 48x40x60 inches
- Pickup date: Next Tuesday (or earliest available)
- Need liftgate delivery
Please send a quote ASAP.
Thanks,
John Smith
Acme Corp
System Requirements
Part 1: Gmail Integration
Requirements:
● Monitor Gmail inbox for new emails (labels/filters optional)
● Read email content and extract metadata (sender, subject, body)
● Send reply emails with PDF attachments
● Mark processed emails appropriately
Implementation Options:
● Gmail API with OAuth2
● IMAP/SMTP (simpler but less robust)
● Any library/SDK you prefer
Part 2: Information Extraction with LLM
Requirements: Extract structured data from unstructured email text:
{
"origin": {
"city": "Los Angeles",
"state": "CA",
"zip": "90021",
"address": "optional"
},
"destination": {
"city": "Chicago",
"state": "IL",
"zip": "60601",
"address": "optional"
},
"cargo": {
"weight_lbs": 800,
"pieces": 2,
"piece_type": "pallets",
"dimensions": {
"length": 48,
"width": 40,
"height": 60,
"unit": "inches"
},
"commodity": "electronics"
},
"special_services": ["liftgate"],
"pickup_date": "2025-11-19",
"additional_notes": ""
}
Must Handle:
● Missing information (partial data)
● Various date formats
● Implicit requirements (e.g., "electronics" implies "needs climate control")
● Multiple units (convert to standard)
Part 3: Quoting API Implementation
Build a mock REST API that calculates freight quotes based on shipment parameters.
API Specification:
Endpoint: POST /api/v1/quote
Example Request Body:
{
"origin_zip": "90021",
"destination_zip": "60601",
"weight_lbs": 800,
"pieces": 2,
"dimensions": {
"length": 48,
"width": 40,
"height": 60
},
"special_services": ["liftgate"],
"pickup_date": "2025-11-19",
"commodity": "electronics"
}
Response:
{
"quote_id": "QT-2025-001234",
"total_cost": 1250.00,
"breakdown": {
"base_rate": 950.00,
"fuel_surcharge": 142.50,
"liftgate_fee": 75.00,
"insurance": 82.50
},
"transit_days": 3,
"equipment_type": "dry_van",
"valid_until": "2025-11-15T23:59:59Z",
"terms": "Payment due upon delivery"
}
Implementation Requirements:
● Use FastAPI, Flask, or any Python web framework
● Include input validation
● Handle edge cases (invalid zips, unrealistic weights)
● Return appropriate error codes
● Include API documentation (Swagger/OpenAPI bonus)
Mock Data:
● Use a simplified zip code database (provide ~50 major US cities)
● Or use real distance API (Google Maps, Mapbox) - bonus points
Part 4: PDF Quote Generation
Generate a professional-looking PDF quote document.
Required Elements:
● Company header/logo (mock freight broker company)
● Quote ID and date
● Origin and destination details
● Shipment details (weight, pieces, dimensions)
● Itemized cost breakdown
● Total price prominently displayed
● Transit time estimate
● Terms and conditions
● Valid until date
● Contact information
Recommended Libraries:
● ReportLab
● WeasyPrint (HTML to PDF)
● FPDF
● Or any PDF generation library
Part 5: Agent Orchestration
Tie everything together in an autonomous agent that:
1. Monitors inbox (polling or webhook)
2. Detects quote request emails
3. Extracts shipment details using LLM
4. Calls quoting API
5. Generates PDF quote
6. Replies to original email with PDF attached
7. Logs the transaction
8. Handles errors gracefully
Error Handling Scenarios:
● Missing critical information → Reply asking for details
● Invalid data (negative weights, etc.) → Reply with error
● API failures → Retry logic or escalate
● Email send failures → Log and retry
Common Pitfalls to Avoid
1. Gmail Authentication: OAuth2 setup can be tricky - start early
2. LLM Prompting: Test extraction with many examples
3. PDF Layout: Basic PDF is fine, don't spend 2 hours on design
4. Over-engineering: Focus on working system first
5. Error Handling: Don't let one failure crash the whole system
Getting Started Tips
1. Start with API: Build and test quote API first
2. Mock early: Use mock emails before Gmail integration
3. Iterate extraction: Test LLM prompts with 10+ examples
4. Simple PDF first: Get basic PDF working, beautify later
5. Test end-to-end early: Don't wait until the end to integrate
Questions to Consider
Your documentation should address:
● How would you scale this to 1000 emails/day?
● What happens if the LLM hallucinates data?
● How do you handle rate limits (Gmail API, LLM API)?
● What's your strategy for improving extraction accuracy?
● How would you add human review for high-value quotes?
Sample Success Flow
1. Email arrives: [login to view URL] @[login to view URL]
Subject: "Need quote for LA to Chicago"
2. System extracts:
✓ Origin: Los Angeles, CA 90021
✓ Destination: Chicago, IL 60601
✓ Weight: 800 lbs
✓ Pieces: 2 pallets
✓ Special: Liftgate
3. API called:
POST /api/v1/quote
→ Quote ID: QT-2025-001234
→ Total: $1,250.00
4. PDF generated:
✓ Professional layout
✓ All details included
✓ Itemized pricing
5. Email sent:
To: [login to view URL] @[login to view URL]
Subject: Re: Need quote for LA to Chicago
Body: "Here's your freight quote..."
Attachment: [login to view URL]
6. Logged:
✓ Transaction recorded
✓ Ready for next email" (client-provided description)