Python Integration Developer for Real Estate AP Automation

Python Integration Developer — Real Estate AP Automation (DocuSign, QuickBooks, Airtable) Project Overview We're a New Jersey-based real estate platform with construction, lending, materials importing, and freight operations across multiple LLCs. We've designed and built ~80% of an accounts payable automation system in Python and need an experienced integration developer to complete three SDK integrations, deploy the system, and run it through end-to-end testing. This is not a greenfield build. The architecture is set, the data model is defined, the code skeleton is written, and all six AI prompts are tested. Your job is to fill in the integration gaps, deploy, and verify everything works end-to-end. What's Already Built We will provide on day one: Complete Python codebase (~3,000 lines, ~12 modules) with full docstrings Detailed Statement of Work / Build Brief System Architecture document Operations Runbook Airtable schema specification (10 tables) DocuSign template specification (2 templates) Six AI prompts (vendor compliance verification, invoice OCR, project inference, GL coding, duplicate detection) Configuration templates (.env, insurance minimums, CS-to-GL mapping with ~80 codes) Deployment guide for hybrid Render + on-prem deployment What You'll Do Three integration completions: DocuSign SDK integration — JWT Grant authentication, send envelopes from templates, download completed envelope documents, verify Connect webhook signatures (skeleton in place; needs docusign-esign SDK calls written out) QuickBooks Online API integration — push approved bills via the python-quickbooks library for our QBO entities (skeleton in place; needs API calls written out and OAuth refresh token management) Email sending — implement the send_email() function via SMTP or a transactional service (Postmark/SendGrid recommended) Deployment: Deploy webhook server to Render (Starter plan) Set up the local runner on a Ubuntu mini-PC in our office (we have the hardware ready) Configure DocuSign Connect, Cloudflare Tunnel (if needed), and cron jobs End-to-end testing: Run real invoices through the system across multiple vendor compliance states Verify approval routing across all four approver tiers (we'll provide test scenarios) Verify QuickBooks Desktop IIF imports and QuickBooks Online API pushes Verify Positive Pay file format with our bank Required Skills Python 3.10+ — solid, idiomatic, async-aware (FastAPI is in the stack) REST API integration experience — you've shipped OAuth2 / JWT Grant integrations before DocuSign API experience — specifically the eSignature REST API, JWT Grant flow, and Connect webhooks. This is the most specialized requirement; please address it specifically in your application. QuickBooks experience — either QuickBooks Online API (for our QBO entities) or QuickBooks Desktop IIF format (for our QB Desktop entities), ideally both Linux (Ubuntu) deployment — comfortable with systemd or supervisor, basic firewall config, cron Git — clean commits, ability to work with our existing branch structure Nice to Have Anthropic API / Claude experience (the system uses Claude for OCR and verification, but the prompts are already written; you don't need prompt engineering skills) Claude Code experience — the codebase was designed to be operated by Claude Code post-deployment; experience here means the maintenance retainer is a natural fit Airtable API experience (we use pyairtable) Slack API for interactive components Real estate, construction, or property management domain familiarity Experience with Cloudflare Tunnel or similar zero-trust networking Engagement & Compensation Hourly, $60-90/hour depending on demonstrated experience with our specific stack 80-hour soft cap — we expect this to land between 60 and 100 hours of work; we'll review at 40 hours and adjust scope/cap if anything has changed Weekly check-ins — Tuesday or Wednesday, 30 minutes, screen-share so we can see what's been done Milestone-gated payments — paid weekly, with each week's work tied to specific deliverables we agree on Monday NDA required before code access — standard mutual NDA; we'll send within 24 hours of selection Working Style We're AI-tooling friendly. Use whatever toolchain makes you fast — Cursor, Copilot, Claude Code, your own setup. Claude Code is a plus for this engagement specifically, because the codebase was designed with Claude-Code-as-orchestration-engine in mind and the bundle includes a docs/claude_code_instructions.md file that lets Claude Code self-orient in the repo. If you already use it, that's a genuine efficiency boost. If you don't, your familiar tools are fine — the work is straightforward Python integration work either way. What we care about more than your tooling: clean commits, clear weekly demos, and code that the next person to touch this repo (potentially Claude Code, potentially us, potentially a different freelancer in 18 months) can read and extend. About Us Mid-sized real estate operating company. Multiple legal entities, professional finance and operations team, in business 8+ years. We work with senior counsel on legal matters and senior accountants on financial systems, so you'll be working alongside (not replacing) competent in-house operators. We pay on time, we communicate clearly, and we don't move goalposts mid-project. How to Apply Please include the following — applications missing these will be skipped: One-paragraph summary of a previous DocuSign API integration you've shipped, including: which auth flow you used (JWT Grant or Authorization Code), how you handled Connect webhooks, and what tripped you up Your QuickBooks experience — Desktop IIF, Online API, or both? Estimate the largest QB integration you've done by transaction volume Your availability — hours per week you can commit, time zone, and earliest possible start date A short answer (3-4 sentences) to: "Our codebase has a function called verify_coi() that calls the Anthropic API to check whether a vendor's Certificate of Insurance meets our requirements. The result of that function determines whether the vendor can be paid. Without changing what the function does, what would you want to know about it before deploying to production?" Question 4 is a thinking question — there's no perfect answer, but it tells us a lot about how you'd approach the work. Please don't include generic cover letter content. We'll read every reply that follows the format above. Estimated timeline: 3-5 weeks from start to production cutover. Time zone preference: US Eastern, but flexible if your overlap is 4+ hours daily. Long-term: A maintenance retainer (~5-10 hours/month) is available after stabilization for the right person.