"Printflow" AI-Powered Print Shop Management System

---

Job Title: Full-Stack Developer to Build “Printflow” – AI-Powered Print Shop Management System

Project Overview:

I am a non-technical business owner looking for an experienced full-stack developer to build a complete management system for a print shop. The system is called Printflow and is based on a detailed architecture PDF that I will provide to the selected freelancer.

Printflow is not just a job tracker. It is three separate but connected domains:

1. Production / Sales – job lifecycle from inquiry to payment.
2. Inventory – materials, stock levels, suppliers, reserved vs consumed quantities.
3. Finance / AP – supplier invoices, payments, purchase costs.

A knowledge graph links all three domains. The system includes:

· Claude (or similar LLM) as a natural-language interface so the owner can ask questions like “Do we have enough material for all scheduled orders?”
· Jev (TypeSafe’s fast decision model) for high-frequency, low-latency decisions such as scoring supplier document confidence and classifying inventory alerts.
· A Telegram bot for the owner to interact with the system.
· Paperless-ngx integration to ingest supplier invoices and delivery notes via OCR.

Scope of Work:

The freelancer will build the complete application in phases. I will provide the architecture PDF and API keys. The freelancer will deliver the full source code. VPS setup, GitHub repository creation, and deployment are NOT part of this job. I will handle those myself.

Phase 1 – Scaffolding

· Create project folder structure (api/, bot/, migrations/, tests/, etc.)
· Create docker-compose.yml and .env.example
· Write a README.md with setup instructions

Phase 2 – Database Schema

· PostgreSQL schema for all three domains
· Tables for jobs, materials, inventory transactions, suppliers, purchase orders, invoices, payments
· Ensure reserved and consumed are separate columns in inventory

Phase 3 – Inventory Service

· Backend API endpoints for materials, stock levels, reserved/consumed quantities
· Logic for supplier deliveries updating inventory
· Reorder level alerts

Phase 4 – Production Service

· Backend API for the 14-step job lifecycle (Inquiry → Paid)
· Material requirements calculation
· Handoff to inventory at the ORDER stage

Phase 5 – Finance / AP Service

· Backend API for supplier invoices, payments, purchase costs
· Link invoices to purchase orders and inventory receipts

Phase 6 – Knowledge Graph

· Create links between jobs, materials, suppliers, and documents
· Use PostgreSQL with a graph table or Neo4j (freelancer’s choice, but must be simple to maintain)

Phase 7 – Claude Tool Endpoints

· Expose safe read-only API endpoints that Claude can call to answer owner questions
· Example: get_open_jobs, get_material_requirements, check_material_availability

Phase 8 – Jev Decision Functions

· Python module using TypeSafe Jev SDK
· Functions: score_supplier_document_confidence, classify_inventory_alert, check_material_availability
· Wire Jev into the Paperless webhook and inventory service

Phase 9 – Telegram Bot

· Build a Telegram bot using python-telegram-bot
· Bot receives owner messages, forwards them to the backend, and returns Claude’s answers
· Include commands for common queries (e.g., /stock, /jobs, /invoices)

Phase 10 – Paperless-ngx Integration

· Webhook handler that receives OCR’d documents from Paperless-ngx
· Extract structured data (supplier, items, quantities, amounts) using an LLM or Jev
· Update inventory and accounts payable accordingly

Phase 11 – Testing & Documentation

· Write unit and integration tests for each service
· Provide clear documentation for a non-technical owner to run the system with docker compose up -d

Deliverables:

· Complete source code (zipped or shared via a file transfer service)
· docker-compose.yml and .env.example
· README.md with step-by-step instructions
· DECISIONS.md listing any assumptions made
· Test suite and instructions to run tests

Required Skills:

· Python 3.12+ and FastAPI
· PostgreSQL and SQLAlchemy (or similar ORM)
· Docker and Docker Compose
· Telegram Bot API (python-telegram-bot)
· Experience integrating LLM APIs (OpenAI, Anthropic Claude, or similar)
· Experience with Paperless-ngx or OCR document processing
· Knowledge graph implementation (PostgreSQL graph tables or Neo4j)
· Ability to follow a detailed architecture PDF
· Willingness to learn and integrate Jev (TypeSafe SDK) – documentation will be provided

Nice to Have:

· Experience with print shop workflows, inventory management, or accounting software
· Familiarity with Jev or other typed decision models
· Ability to explain technical concepts in plain English to a non-technical client

What I Will Provide:

· The full architecture PDF (Printflow_Architecture.pdf)
· API keys for Claude (Anthropic) and Jev (TypeSafe)
· Telegram bot token
· Paperless-ngx instance (or instructions to set one up)
· Prompt feedback and clarification throughout the project

Not Included in This Job:

· VPS setup or server configuration
· GitHub repository creation or management
· Deployment to production
· Ongoing maintenance after delivery

Timeline:

Estimated 4–8 weeks depending on experience and availability. Please provide your estimated timeline in your proposal.

Budget:

Open to proposals. Please provide a detailed quote with a breakdown of hours per phase.

How to Apply:

Please include:

1. Examples of similar full-stack projects you have built (especially with AI/LLM integration).
2. Your experience with Python, FastAPI, PostgreSQL, Docker, and Telegram bots.
3. Your approach to building this system in phases.
4. Your estimated timeline and cost.
5. Any questions you have about the project.

I am looking for a reliable developer who can communicate clearly and deliver high-quality work. I am not a coder, so I need someone who can guide me through testing and understanding the final product