AI-Driven Pre-sales Automation Platform

We are building an AI-powered pre-sales automation platform to replace the manual workload of our engineering team across 15 ELV disciplines.

The platform allows our engineers to upload a client RFQ (Bill of Quantities in Excel or PDF), and the system automatically produces a fully priced, formatted Excel quotation — with part number mapping, pricing calculations, BOQ validation, and an internal analysis sheet — in under 2 hours.

This is Phase 1 of a 4-phase project. Phase 1 scope is the Quotation Engine only.

**Phase 1 Scope (This Contract)**

You will build a complete web application with the following:

**Backend (Python FastAPI)**
- File upload API: accepts Excel (.xlsx) and PDF client BOQ files
- BOQ parser: extracts device descriptions and quantities from Excel and PDF
- Spec parser: extracts brand/vendor name and applicable standards from PDF specification
- AI device mapper: calls Anthropic Claude API to match each device description to our part number catalog
- Pricing engine: applies a formula chain (list price × multiplier × exchange rate → cost → sell price) using parameters stored in database
- BOQ validator: calls Claude API to review the mapped BOQ and flag missing items, unusual quantities, and scope gaps
- Excel output generator: produces a 3-sheet Excel workbook (Front Page, Material BOQ, Analysis Sheet) in our exact company format using openpyxl

**Frontend (React.js + Tailwind CSS)**
- Login page (Supabase Auth)
- Project creation form: system type, client, reference number, brand, pricing parameters
- File upload UI with progress indicator
- BOQ preview table: shows AI-mapped items, quantities (editable), unit prices — engineer can override before generating
- AI flags panel: shows validator output for engineer review
- Generate & Download button: produces final Excel file
- Project history page: searchable list of past quotations

**Database & Infrastructure**
- PostgreSQL via Supabase: price lists, vendor parameters, projects, users
- File storage via Supabase Storage
- Frontend deployed to Vercel
- Backend deployed to Railway
- Custom domain configuration
- GitHub private repository with CI/CD

---

**Tech Stack Required**

- Backend: Python 3.11 + FastAPI
- Frontend: React.js + Tailwind CSS
- Database: PostgreSQL + Supabase
- File processing: pdfplumber, openpyxl, python-docx
- AI integration: Anthropic Claude API (claude-sonnet)
- Hosting: Vercel (frontend) + Railway (backend)
- Version control: GitHub

---

**Timeline**

12 weeks total:
- Week 1: Foundation setup (database schema, environment, CI/CD pipeline)
- Weeks 2–6: Backend development (parser, AI mapper, pricing engine, output generator)
- Weeks 4–8: Frontend development (overlapping with backend)
- Weeks 9–12: Testing with real projects, bug fixes, deployment, handover

---

**What We Provide**

- Detailed project charter with all specifications
- Exact Excel output template (column structure, formula logic, styling)
- Vendor pricing parameter table (all pricing rules)
- Product price list data (ready to load into database)
- 5 real pilot projects for testing
- Anthropic API key
- Supabase, GitHub, and Vercel accounts (already created)
- Weekly check-in calls (30 minutes, every Monday)
- Fast feedback and decision-making

---

**Requirements**

- Proven experience with Python FastAPI and React.js
- Experience with PDF and Excel parsing and generation
- Experience integrating OpenAI or Anthropic API in a production application
- Experience with Supabase or similar (PostgreSQL + Auth + Storage)
- Experience deploying to Vercel and Railway or equivalent
- Portfolio showing at least one document processing or data-heavy web application
- Strong English communication
- Available for the full 12-week duration
- Able to start within 1 week

---

**Contract Type**

Fixed price for Phase 1. depending on experience and proposal quality. Ongoing work available for Phases 2–4 (material submittal engine, engineering calculations, diagram generation) for the right candidate.

---

**To Apply**

Please include in your proposal:
1. A link to at least one project where you processed PDF or Excel files and generated structured output
2. Confirmation you have worked with Claude API or OpenAI API
3. Your fixed-price bid and estimated delivery timeline
4. Your availability to start

Short proposals without portfolio links will not be considered.