Full-Stack Software Engineer — Greenfield Receivables Platform (Project 47)

Full-Stack Software Engineer — Greenfield Receivables Platform (Project 47) Company: Confidential — a Connecticut receivables & civil-litigation firm Location: Old Saybrook, Connecticut — [remote-friendly (US) / hybrid — set this] Type: [Full-time / contract — set this] About us We are a small, owner-operated Connecticut firm in receivables and civil-litigation enforcement. We acquire civil judgments and pursue lawful post-judgment enforcement — wage executions and bank garnishments — under state and federal law. We're profitable, lean, and serious about doing this work correctly and compliantly. About the project Project 47 is a ground-up replacement for our system of record — today a decades-old Windows desktop application that everything runs through. We're building a modern, browser-native platform that manages debtor accounts, portfolios, financial ledgers, and the document machine that drives the business: court forms, letters, motions, and the pipelines that ingest and produce them. The hard early work is done. The court-forms module and the letters/motions module are built and validated as working prototypes, the data model is mapped against the legacy schema, and the app shell is locked. We're now moving from validated mockups into full production — and we want an engineer to own that transition. The role We're hiring a mid-to-senior full-stack engineer to take Project 47 from prototype to production. You'll own features end to end — data model, API, and UI — and work directly with the owner and our existing developer. This is high-ownership work on a small team: what you build ships, and you'll see the business using it within days, not quarters. What you'll work on Build the production application — React/TypeScript front end, Hono APIs on AWS Lambda, PostgreSQL, and S3 — turning a set of locked prototypes into a real, deployed system. Design the data model for accounts, portfolios, transactions and ledgers, co-signers, and document records, migrating cleanly off a legacy schema without mangling decades of keyed history. Build document generation — court-form and letter/motion rendering driven by a merge-field/template system, with real print fidelity, multi-page handling, and signature/notary apparatus. Develop serverless document pipelines for PDF and text extraction, classification, and structured output. Several already run on Lambda and use the Anthropic (Claude) API for extraction and fallback parsing. Integrate external data and services — court/e-filing data, print-and-mail vendors, employment and asset verification, and payment processing. Engineer for compliance and auditability from the start: CFPB Reg F-aware workflows, audit trails, and field-level controls are first-class requirements, not afterthoughts. Tech stack Frontend: React, TypeScript Backend: AWS Lambda (Hono), Node / TypeScript Data: PostgreSQL, S3 AI: Anthropic Claude API (in production, in the document pipelines) Tooling: Git, automated validation and testing What we're looking for Strong full-stack experience with React + TypeScript and a modern backend. Solid relational data modeling — you can design a normalized PostgreSQL schema and reason confidently about migrations. Hands-on AWS, especially serverless (Lambda, S3, IAM); you're comfortable owning the infrastructure as part of the job. A track record shipping document-heavy or workflow-heavy applications — generation, parsing, PDFs, templating, batch processing. The ability to work autonomously, turn fuzzy business requirements into shipped features, and communicate clearly in writing. Nice to have Production experience integrating LLMs — structured extraction, prompt design, evaluation. Background in legal-tech, fintech, collections/receivables, or another regulated, compliance-heavy domain. Familiarity with PDF tooling (pdfplumber / pypdf, form filling), court/e-filing data, or print/mail pipelines. Python for data-pipeline work. Why this role is worth your time Greenfield. You're not nursing legacy spaghetti — you're replacing it. Real impact, fast. Small team, a direct line to the owner, and work that goes into daily use immediately. AI that's actually in production — not a demo, not a science project. Careful engineering that matters. Correctness and compliance are the point here, and good work is visible and valued.